Quotes and Jokes

Machine learning is statistics minus any checking of models and assumptions. (Brian Ripley)

The problem with machine learning is that the machine does all the learning. (Unknown)

We have more data than ever, more good data than ever, a lower proportion of data that are good, a lack of strategic thinking about what data are needed to answer questions of interest, sub-optimal analysis of data, and an occasional tendency to do research that should not be done. (Frank Harrell)

In theory, there is no difference between theory and practice. In practice, there is. (Benjamin Brewster)

The difference between practice and theory is greater in practice than in theory. (corollary to the above)

Statisticians tend to forget that there is a rather important stage between planning a clinical trial and analysing it called ‘doing a trial’. (Peter Armitage)

Applying the theory of personal probability naively, one quickly comes to the conclusion that randomization is without value for statistics. This conclusion does not sound right; and it is not right. Closer examination of the road to this unintentional conclusion does lead to new insights into the role and limitations of randomization but does by no means deprive randomization of its important function in statistics (L. J. Savage)

Though we all feel sure that randomization is an important invention, the theory of subjective probability reminds us that we have not fully understood randomization. The need for randomization presumably lies in the imperfection of actual people and, perhaps, in the fact that more than one person is ordinarily concerned with an investigation. (L. J. Savage)

We all know that art is not truth. Art is a lie that makes us realize truth, at least the truth that is given us to understand. The artist must know the manner whereby to convince others of the truthfulness of his lies. If he only shows in his work that he has searched, and re-searched, for the way to put over his lies, he would never accomplish any thing. (Pablo Picasso)

The mathematics are not there for the joy of the analyst but because they are essential to the solution. (Karl Pearson)

Whenever there is a simple error that most laymen fall for, there is always a slightly more sophisticated version of the same problem that experts fall for. (Amos Tversky)

Bayes = Distributional + prior assumption. Frequentist = Distributional + sampling dist assumption. You don’t need a prior to be ‘true’, you need it to be defendable. “Given this prior uncertainty, what do the data suggest?” Can you defend the existence of a sampling distribution? (Stephen Martin)

People don’t go around introducing you to their ex-wives.” (on why model improvement doesn’t make it into papers) (Andrew Gelman)

Statistics is said to be the science of defaults. One of our challenges is to defaultize things (Andrew Gelman)

This guy comes to me and says ‘I have prior information and data, and I’d like to combine them, and I heard Bayes is a good way to do that.’ Well, that’s as good as it gets! Normally you want to do Bayes but they won’t let you because they’re like [in stupid voice] ‘ugh, it’s subjective, I’m not allowed to, it’s subjective.’ But here this guy is saying ‘I have prior information and data and I want to combine them’ ! I’m like, ’I can do that! I was trained to do that! (Andrew Gelman)

All models are wrong, but some are useful. (George E. P. Box)

In statistics it’s enough for our results to be cool. In psychology they’re supposed to becorrect. In economics they’re supposed to be correct and consistent with your ideology. (Andrew Gelman)

As you know from teaching introductory statistics, 30 is infinity. (Andrew Gelman)

Suppose there’s someone you want to get to know better, but you have to talk to all herfriends too. They’re like the nuisance parameters (Andrew Gelman)

Exchangeability is a function not just of reality, but of the information you have. (Andrew Gelman)

Inference is normal science. Model-checking is revolutionary science. (Andrew Gelman)

(Someone asks how to compare models. Gelman writes in giant letters: OUT OF SAMPLE PREDICTION ERROR.) “So there’s that.” (Andrew Gelman)

The full name of theoretical statistics is the theory of applied statistics (Andrew Gelman)

An approximate answer to the right problem is worth a good deal more than an exact answer to an approximate problem (John Tukey)

To call in the statistician after the experiment is done may be no more than asking him to perform a post-mortem examination: he may be able to say what the experiment died of. (RA Fisher)

In God we trust. All others must bring data. (Deming)

Statisticians, like artists, have the bad habit of falling in love with their models. (GEP Box)

Prediction is very difficult, especially about the future. (Bohr)

If you torture the data enough, nature will always confess. (Ron Coase)

A big computer, a complex algorithm and a long time does not equal science. (Rob Gentleman)

The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data (John Tukey)

There are no routine statistical questions, only questionable statistical routines. (DR Cox)

Statistics - A subject which most statisticians find difficult but which many physicians are experts on. (Stephen S. Senn)

Strange events permit themselves the luxury of occurring (Charlie Chan)

The best thing about being a statistician is that you get to play in everyone’s backyard. (John Tukey)

Absence of evidence is not evidence of absence (Martin Rees)

It’s easy to lie with statistics; it is easier to lie without them. (Fred Mosteller)

For a Bayesian, the problem with the “bias” concept is that is conditional on the true parameter value. But you don’t know the true parameter value. There’s no particular virtue in unbiasedness. (Andrew Gelman)

(Translated) Everybody believes in the exponential law of errors [i.e., the Normal distribution]: the experimenters, because they think it can be proved by mathematics; and the mathematicians, because they believe it has been established by observation. (Poincare)

… surely, God loves the .06 nearly as much as the .05. Can there be any doubt that God views the strength of evidence for or against the null as a fairly continuous function of the magnitude of p? (Rosnow and Rosenthal)

On two occasions I have been asked [by members of Parliament], ‘Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?’ I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question. (Babbage)

The subjectivist (i.e. Bayesian) states his judgements, whereas the objectivist sweeps them under the carpet by calling assumptions knowledge, and he basks in the glorious objectivity of science. (IJ Good)

Figures don’t lie, but liars do figure (Mark Twain)

If your experiment needs statistics, you ought to have done a better experiment. (Ernest Rutherford)

To find out what happens when you change something, it is necessary to change it. (Box, Hunter, and Hunter)

We are drowning in information and starving for knowledge. (Rutherford Roger)

The greatest value of a picture is when it forces us to notice what we never expected to see. (John Tukey)

The statistician cannot evade the responsibility for understanding the process he applies or recommends. (RA Fisher)

Data do not speak for themselves - they need context, and they need skeptical evaluation (Allen Wilcox)

A wise man, therefore, proportions his belief to the evidenceA wise man, therefore, proportions his belief to the evidence (Hume)

My thesis is simply this: probability does not exist. (Bruno de Finetti)

Taking a model too seriously is really just another way of not taking it seriously at all. (Andrew Gelman)

What the use of a p-value implies, therefore, is that a hypothesis that may be true may be rejected because it has not predicted observable results that have not occurred. (Jeffreys)

These days the statistician is often asked such questions as “Are you a Bayesian?” “Are you a frequentist?” “Are you a data analyst?” “Are you a designer of experiments?”. I will argue that the appropriate answer to ALL of these questions can be (and preferably should be) “yes”, and that we can see why this is so if we consider the scientific context for what statisticians do. (GEP Box)

Uncertainty is a personal matter; it is not the uncertainty but your uncertainty. (Dennis Lindley)

R Quotes/Wisdom

This page is the complete output of the fortunes package, in a convenient hypertext format :)

library(fortunes)
dat <- read.fortunes()
for (i in seq_len(nrow(dat))) print(fortune(i))
## 
## Okay, let's stand up and be counted: who has been writing diamond graph code?
## Mine's 60 lines.
##    -- Barry Rowlingson (in a discussion about the patent for diamond graphs)
##       R-help (August 2003)
## 
## 
## Bug, undocumented behaviour, feature? I don't know. It all seems to work in
## 1.6.0, so everyone should downgrade now... :)
##    -- Barry Rowlingson
##       R-help (July 2003)
## 
## 
## I'm always thrilled when people discover what lexical scoping really means.
##    -- Robert Gentleman
##       Statistical Computing 2003, Reisensburg (June 2003)
## 
## 
## My institution has a particularly diabolical policy on intellectual property,
## especially on software.
##    -- Ross Ihaka
##       R-help (August 2003)
## 
## 
## If you imagine that this pen is Trellis, then Lattice is not this pen.
##    -- Paul Murrell (on the difference of Lattice (which eventually was called
##       grid) and Trellis)
##       DSC 2001, Wien (March 2001)
## 
## 
## It is good to look around at what is available, but one shouldn't miss what is
## under one's nose either.
##    -- Robert Gentleman
##       bioconductor (September 2003)
## 
## 
## What we have is nice, but we need something very different.
##    -- Robert Gentleman
##       Statistical Computing 2003, Reisensburg (June 2003)
## 
## 
## Readable, obvious, maintainable, 'portable' for some value of 'portable'...
##    -- Barry Rowlingson (suggesting some function)
##       R-help (September 2003)
## 
## 
## 'They' did write documentation that told you that Perl was needed, but 'they'
## can't read it for you.
##    -- Brian D. Ripley
##       R-help (February 2001)
## 
## 
## Overall, SAS is about 11 years behind R and S-Plus in statistical capabilities
## (last year it was about 10 years behind) in my estimation.
##    -- Frank Harrell (SAS User, 1969-1991)
##       R-help (September 2003)
## 
## 
## I want a budget.
##    -- Peter Dalgaard (in a talk about Tcl/Tk, meaning to say 'I want a button
##       widget')
##       gR 2003, Aalborg (September 2003)
## 
## 
## You can be maximally lazy, but still be efficient.
##    -- Kevin Murphy (describing the implementation of an algorithm)
##       gR 2003, Aalborg (September 2003)
## 
## 
## Andrew Thomas: ...and if something goes wrong here it is probably not WinBUGS
## since that has been running for more than 10 years...
## Peter Green (from the back): ... and it still hasn't converged!
##    -- Andrew Thomas and Peter Green (during the talk about 'BRugs')
##       gR 2003, Aalborg (September 2003)
## 
## 
## This is all documented in TFM. Those who WTFM don't want to have to WTFM again
## on the mailing list. RTFM.
##    -- Barry Rowlingson
##       R-help (October 2003)
## 
## 
## It really is hard to anticipate just how silly users can be.
##    -- Brian D. Ripley
##       R-devel (October 2003)
## 
## 
## So apparently you wish to report as a bug the fact that R 1.8.0 is different
## from R 1.4.0.
##    -- Douglas Bates
##       R-devel (October 2003)
## 
## 
## I plan to fix this - the report is just in case I forget or get hit by a bus.
##    -- Ross Ihaka (reporting a bug in persp)
##       R-help (October 2003)
## 
## 
## Thomas W. Blackwell: I just discovered to my surprise that I cannot define a
## function with an argument named 'break' or 'while'!
## Brian D. Ripley: No, they are reserved names. [...] Most programming languages
## have reserved words, so I am surprised you are surprised.
##    -- Thomas W. Blackwell and Brian D. Ripley
##       R-help (November 2003)
## 
## 
## R has changed quite a lot recently, and older preconceptions do need to be
## checked against current information.
##    -- Brian D. Ripley
##       R-help (November 2003)
## 
## 
## But a difference of a few minutes means that this is well under 20% of the
## total time unless your statistical analysis is very much speedier than mine.
##    -- Brian D. Ripley
##       R-devel (November 2003)
## 
## 
## I seem to recall that we were targetting 512k Macintoshes. In our dreams we
## might have seen 16Mb Sun.
##    -- Ross Ihaka (in reply to the question whether R&R thought when they
##       started out that they would see R using 16G memory on a dual Opteron
##       computer)
##       R-help (November 2003)
## 
## 
## I quit using SAS in 1991 because my productivity jumped at least 20% within one
## month of using S-Plus.
##    -- Frank Harrell
##       R-help (November 2003)
## 
## 
## When a Certain Guru rips strips off people (God knows he's done it to me often
## enough) on this list, there's a damned good reason for it.
##    -- Rolf Turner (in a discussion about whether a friendly mailing list with
##       more 'customer service' attitude than R-help was needed)
##       R-help (December 2003)
## 
## 
## Ah, so that's why my report of a bug in the RCheapViagra package didn't get
## through....
##    -- Barry Rowlingson (after Peter Dalgaard announced a new spam filter for
##       R-bugs)
##       R-devel (January 2004)
## 
## 
## Henrik Bengtsson: Is there a way to turn off the (annoying) beep that occurs
## when one calls the locator() command and clicks the mouse? [...]
## Brian D. Ripley: It's a feature of the windows() device. Windows is fond of
## beeping, and I just mute the sound.
##    -- Henrik Bengtsson and Brian D. Ripley
##       R-help (July 2001)
## 
## 
## Brian D. Ripley: Add to package utils in R-devel, after correction. I was
## surprised you had fallen into the 1:0 trap.
## Patrick Burns: I'm surprised too -- good catch.
##    -- Brian D. Ripley and Patrick Burns (after adding Patrick Burns' head() to
##       the utils package)
##       R-devel (January 2004)
## 
## 
## As to whether you can do a Lilliefors test for several groups, that depends
## entirely on your ability to understand what the underlying question would be
## (see Adams D 1979).
##    -- Knut M. Wittkowski
##       R-help (February 2004)
## 
## 
## Jim Gustafsson: I would like to put my SAS-code into R. Could I do that, if
## yes, how?
## Frank Harrell: Just reverse the procedure you use when you put R code into SAS.
## ;)
##    -- Jim Gustafsson and Frank Harrell
##       R-help (February 2004)
## 
## 
## Is this English or American (you know, the language referred to in the USA as
## 'English')? [...] As an English English speaker, my sense and my employer's
## dictionary both suggest doubling here.
##    -- Brian D. Ripley (in a discussion whether it is 'subsetable' or
##       'subsettable')
##       R-devel (February 2004)
## 
## 
## As Brian says, there's nothing in the design that lets you do this, but then
## since you have the source there's nothing on the planet to stop you doing this.
## The design is not a law :)
##    -- Barry Rowlingson (answering the question whether the mouse position can
##       be determined without any click)
##       R-help (February 2004)
## 
## 
## Barry Rowlingson: Your grid above has 8*6 = 42 points.
## (That was a subtle Hitchhikers Guide To The Galaxy reference there, honest, and
## not a stupid dumb multiplication mistake on my part after working four 18-hour
## days on the trot...)
## Peter Dalgaard: [...] Don't panic, just throw yourself at the ground and miss.
##    -- Barry Rowlingson and Peter Dalgaard
##       R-help (March 2004)
## 
## 
## I have mentioned several times on this list that I'm in the process of
## developing a new and wonderful implementation of lme and I would prefer to
## continue working on that rather than modifying old-style code.
##    -- Douglas Bates
##       R-help (March 2004)
## 
## 
## Ivo Welch: Thanks. I will put in a suggestion that the docs refer to q() in
## 'see also' for 'stop'.
## Brian D. Ripley: I don't think anyone else is confusing 'exit' with 'stop',
## though. I hope you don't when driving ....
##    -- Ivo Welch and Brian D. Ripley
##       R-help (March 2004)
## 
## 
## Before we get too carried away with this thread could you all please consider
## how the sd function calculates its result? [...] I'll tell you, it takes the
## square root of the variance. How is the variance calculated for a numeric
## vector? First you calculate the mean *using floating point arithmetic* in which
## it is not necessarily true that N * k / N == k [...] Most of those tests [for
## numerical accuracy] end in a check using the all.equal function which checks if
## the relative difference is less than a threshold. That's about the best that
## you can do with floating point arithmetic.
## Here endeth the sermon.
##    -- Douglas Bates (after a discussion why sd(rep(0.001, 15)) is not
##       necessarily exactly 0)
##       R-help (March 2004)
## 
## 
## The keyboard is the standard Italian layout, which is missing the ~ (tilde)
## key. [...] Can anyone advice how to produce the ~ symbol, short of a copy/paste
## from MS Word?
##    -- Federico Calboli
##       R-help (March 2004)
## 
## 
## I'd like to prefix all these solutions with 'Here's how to do it, but don't
## actually do it you crazy fool'. It's on a par with redefining pi, or redefining
## '+'. And then redefining '<-'. These techniques have their proper place, and
## that would be in the currently non-existent obfuscated R contest.
## No, the R-ish (iRish?) way is to index vectors from 1. That's what the R gods
## intended!
##    -- Barry Rowlingson (in a discussion how vectors in R could be indexed
##       starting from 0)
##       R-help (March 2004)
## 
## 
## I think that it's generally a good idea not to resist the most natural way of
## programming in R.
##    -- John Fox
##       R-help (March 2004)
## 
## 
## For almost 40 years SAS has been the primary tool for statisticians worldwide
## and its easy-to-learn syntax, unsurpassed graphical system, powerful macro
## language and recent graphical user interfaces have made SAS the number one
## statistical software choice for both beginners and advanced users.
##    -- Rolf Poalis, Biostatistics Denmark (announcement of the SAS to R parser
##       sas2R)
##       R-help (April 1, 2004)
## 
## 
## Sorting is a wonderful topic! Especially because you can discuss different
## fundamental ideas like brute force, divide and conquer, and questions of
## efficiency, tradeoffs of space and time, etc.
##    -- Peter Wolf
##       R-help (April 2004)
## 
## 
## Douglas Bates: If you really want to be cautious you could use an octal
## representation like sep="\\007" to get a character that is very unlikely to
## occur in a factor level.
## Ed L. Cashin: I definitely want to be cautious. Instead of the bell character I
## think I'll use the field separator character, "\\034", just because this is the
## first time I've been able to use it for it's intended purpose! ;)
## Douglas Bates: Yes, but with "\\034" you don't get to make obscure James Bond
## references :-)
##    -- Douglas Bates and Ed L. Cashin
##       R-help (April 2004)
## 
## 
## Bret Collier: I hope this is not a uniformed question, but I am a little lost.
## Peter Dalgaard: Don't worry, they all look alike... ;-)
##    -- Bret Collier and Peter Dalgaard
##       R-help (April 2004)
## 
## 
## I was actually reading it with some curiosity as to how they managed to find 5
## locations that were close to everyone on R-help...
##    -- Peter Dalgaard (after XL Solutions announced courses 'R/Splus Programming
##       Techniques, @ 5 locations near you!')
##       R-help (April 2004)
## 
## 
## My preference goes with the numbering scheme attributed to a tribe on some
## island in the Pacific which consists of a 'factor' with four levels: 'one',
## 'two', 'three', and 'lots'. Hence, I'd go with 'lots of R users'.
##    -- Dirk Eddelbuettel (in a discussion about trying to estimate the number of
##       R users)
##       R-help (April 2004)
## 
## 
## The Huli of Papua New Guinea use '15' to mean a very large number and '15 times
## 15 samting (something)' to mean something close to infinity.
##    -- David Whiting (in a discussion about trying to estimate the number of R
##       users)
##       R-help (April 2004)
## 
## 
## It seems to me against the spirit of Open Source software to attempt to monitor
## distribution. We could ask R to 'call home' on first use (in the way e.g. pine
## does) but I suspect many users would find that objectionable.
##    -- Brian D. Ripley (in a discussion about trying to estimate the number of R
##       users)
##       R-help (April 2004)
## 
## 
## Maybe you should contact Microsoft, and get them to modify their tools so they
## work on the R sources?
##    -- Duncan Murdoch (answering the question if there is a way compiling the R
##       sources using VisualC++)
##       R-help (April 2004)
## 
## 
## Seldom are prizes, credit, and gratitude given, else Brian would be drowning in
## them.
##    -- Anthony Rossini (about the merits of implementing software)
##       R-help (May 2004)
## 
## 
## Release 1.0.0
## (silence)
## Wow! Thank you! [...] If I am allowed to ask just one question today: How do
## you fit 48 hours of coding in an ordinary day? Any hints will be appreciated
## ... :-)
##    -- Detlef Steuer (on 2000-02-29)
##       R-help (February 2000)
## 
## 
## If you give people a linear model function you give them something dangerous.
##    -- John Fox
##       useR! 2004, Vienna (May 2004)
## 
## 
## To paraphrase provocatively, 'machine learning is statistics minus any checking
## of models and assumptions'.
##    -- Brian D. Ripley (about the difference between machine learning and
##       statistics)
##       useR! 2004, Vienna (May 2004)
## 
## 
## The documentation level of R is already much higher than average for open
## source software and even than some commercial packages (esp. SPSS is notorious
## for its attitude of "You want to do one of these things. If you don't
## understand what the output means, click help and we'll pop up five lines of
## mumbo-jumbo that you're not going to understand either.")
##    -- Peter Dalgaard
##       R-help (April 2002)
## 
## 
## Can one be a good data analyst without being a half-good programmer? The short
## answer to that is, 'No.' The long answer to that is, 'No.'
##    -- Frank Harrell
##       1999 S-PLUS User Conference, New Orleans (October 1999)
## 
## 
## For the S system, which has forever altered the way people analyze, visualize,
## and manipulate data .... S is an elegant, widely accepted, and enduring
## software system, with conceptual integrity, thanks to the insight, taste, and
## effort of John Chambers.
##    -- Association for Computing Machinery
##       ACM/Software System Award citation (1998)
## 
## 
## Some of us feel that type III sum of squares and so-called ls-means are
## statistical nonsense which should have been left in SAS.
##    -- Brian D. Ripley
##       s-news (May 1999)
## 
## 
## I'm really curious to know why the "two types" of sum of squares are called
## "Type I" and "Type III"! This is a very common misconception, particularly
## among SAS users who have been fed this nonsense quite often for all their
## professional lives. Fortunately the reality is much simpler. There is, by any
## sensible reckoning, only ONE type of sum of squares, and it always represents
## an improvement sum of squares of the outer (or alternative) model over the
## inner (or null hypothesis) model. What the SAS highly dubious classification of
## sums of squares does is to encourage users to concentrate on the null
## hypothesis model and to forget about the alternative. This is always a very bad
## idea and not surprisingly it can lead to nonsensical tests, as in the test it
## provides for main effects "even in the presence of interactions", something
## which beggars definition, let alone belief.
##    -- Bill Venables
##       R-help (November 2000)
## 
## 
## I was profoundly disappointed when I saw that S-PLUS 4.5 now provides "Type
## III" sums of squares as a routine option for the summary method for aov
## objects. I note that it is not yet available for multistratum models, although
## this has all the hallmarks of an oversight (that is, a bug) rather than common
## sense seeing the light of day. When the decision was being taken of whether to
## include this feature, "because the FDA requires it" a few of my colleagues and
## I were consulted and our reply was unhesitatingly a clear and unequivocal "No",
## but it seems the FDA and SAS speak louder and we were clearly outvoted.
##    -- Bill Venables
##       'Exegeses on Linear Models' paper (May 2000)
## 
## 
## SAS seems to be to statistical computing what Microsoft is to personal
## computing.
##    -- Bill Venables
##       'Exegeses on Linear Models' paper (May 2000)
## 
## 
## Zhu Wang: I am trying to create a library which uses some Fortran source files
## [...]
## Douglas Bates: Someone named Martin Maechler will shortly be sending you email
## regarding the distinction between 'library' and 'package' :-)
##    -- Zhu Wang and Douglas Bates
##       R-help (May 2004)
## 
## 
## Let's not kid ourselves: the most widely used piece of software for statistics
## is Excel.
##    -- Brian D. Ripley ('Statistical Methods Need Software: A View of
##       Statistical Computing')
##       Opening lecture RSS 2002, Plymouth (September 2002)
## 
## 
## John Miller: How do I prevent google search to post my questions asked here??
## Martin Maechler: you don't: R-help is famous and celebrity can't be gotten rid
## off ;-)
##    -- John Miller and Martin Maechler
##       R-help (June 2004)
## 
## 
## I use R. My company benefits from it. My clients benefit from it.
## ...and I sleep just fine (when I do sleep)... :-)
##    -- Marc Schwartz, Medanalytics (about the 'costs' of free software)
##       R-help (June 2004)
## 
## 
## Please bear with a poor newbee, who might be doing everything backwards (I was
## brought up in pure math).
##    -- Thomas Poulsen
##       R-help (May 2004)
## 
## 
## 1st mail: One of the reasons that SAM is popular, is that it is popular (i.e.
## since everyone has heard of it, it makes reviewers happy). So, it would be nice
## to be able to point to publications in good journals so that reviewers will be
## comfortable. (I personally, am quite comfortable with SAM).
## 2nd mail: Oops, must have been a Freudian slip. Actually, I am not perfectly
## comfortable with SAM. But I am quite comfortable with limma.
##    -- Naomi Altman (answering the question whether anyone had published data
##       using limma for microarray analysis)
##       bioconductor (June 2004)
## 
## 
## However, if you want to do this at all efficiently for a data frame, start with
## my solution not Uwe's [...]
##    -- Brian D. Ripley (in a second reply to a question related to data frame
##       manipulations)
##       R-help (June 2004)
## 
## 
## /* some real insanity to keep Duncan sane */
##    -- Luke Tierney
##       in src/main/context.c (June 2001)
## 
## 
## Roger D. Peng: I don't think anyone actually believes that R is designed to
## make *everyone* happy. For me, R does about 99% of the things I need to do, but
## sadly, when I need to order a pizza, I still have to pick up the telephone.
## Douglas Bates: There are several chains of pizzerias in the U.S. that provide
## for Internet-based ordering (e.g. www.papajohnsonline.com) so, with the
## Internet modules in R, it's only a matter of time before you will have a
## pizza-ordering function available.
## Brian D. Ripley: Indeed, the GraphApp toolkit (used for the RGui interface
## under R for Windows, but Guido forgot to include it) provides one (for use in
## Sydney, Australia, we presume as that is where the GraphApp author hails from).
## Alternatively, a Padovian has no need of ordering pizzas with both home and
## neighbourhood restaurants ....
##    -- Roger D. Peng, Douglas Bates, and Brian D. Ripley
##       R-help (June 2004)
## 
## 
## Uwe Ligges: Please stop sending unsensible bug reports! Those have to be
## handled manually in the bug repository!
## Barry Rowlingson: Really? They seem to be being handled automatically and
## frighteningly well by the Uwe-bot at the moment. Congratulations, you've passed
## the Turing Test.
##    -- Uwe Ligges and Barry Rowlingson (after several nonsensical mails to
##       R-bugs)
##       R-devel (July 2004)
## 
## 
## Soon, they'll be speaking R on the subway.
##    -- Michael Rennie (giving 'Kudos to the R support team')
##       R-help (July 2004)
## 
## 
## Ted Harding: But you can also do these with 'any' and 'all', e.g. any(v==TRUE).
## Thomas Lumley: or any( (v==TRUE)==TRUE), or any( ((v==TRUE)==TRUE)==TRUE)...
## Or, perhaps, any(v). Lewis Carroll wrote a nice piece on this theme.
##    -- Ted Harding and Thomas Lumley (about implementing an 'or' of a logical
##       vector)
##       R-help (August 2004)
## 
## 
## Benjamin Lloyd-Hughes: Has anyone had any joy getting the rgdal package to
## compile under windows?
## Roger Bivand: The closest anyone has got so far is Hisaji Ono, who used MSYS
## (http://www.mingw.org/) to build PROJ.4 and GDAL (GDAL depends on PROJ.4,
## PROJ.4 needs a PATH to metadata files for projection and transformation), and
## then hand-pasted the paths to the GDAL headers and library into src/Makevars,
## running Rcmd INSTALL rgdal at the Windows command prompt as usual. All of this
## can be repeated, but is not portable, and does not suit the very valuable
## standard binary package build system for Windows. Roughly: [points 1 to 5 etc
## omitted]
## Barry Rowlingson: At some point the complexity of installing things like this
## for Windows will cross the complexity of installing Linux... (PS excepting
## live-Linux installs like Knoppix)
##    -- Benjamin Lloyd-Hughes, Roger Bivand, and Barry Rowlingson
##       R-help (August 2004)
## 
## 
## So good advice here is: Beware of good advice about this. (Of course, I may
## just be an outlier ...)
##    -- Berton Gunter (replying to the question what the best way to detect an
##       outlier is)
##       R-help (September 2004)
## 
## 
## Gabor Grothendieck: Congratulations to the R team and all involved for reaching
## the 2.0.0 milestone. The progress of R is truly astounding.
## Barry Rowlingson: A milestone is something that tells you how far it is to
## where you are going. With R-2.0.0, have we arrived?
## Peter Dalgaard: Depending on direction, milestones might only be telling you
## how far you've gone. And you may even be walking in circles around the
## target/origin...
##    -- Gabor Grothendieck, Barry Rowlingson, and Peter Dalgaard
##       R-help (October 2004)
## 
## 
## So it ***do not refuse*** to read the data. I do not expect some dwarf climbs
## out from your computer and says he will not read your data. Or is he?
## ?read.table gives you more details about how to read some data.
##    -- Petr Pikal (after a post that said read.table() refuses to read
##       unbalanced data)
##       R-help (October 2004)
## 
## 
## These road distances seem to be very non-Euclidean indeed (even non-metric). It
## seems to be 2282km from Athens to Milan if you go directly, but if you go via
## Rome it is only 1403km. I thought that World is non-Euclidean, but not that
## obviously.
##    -- Jari Oksanen (on the eurodist data)
##       R-help (October 2004)
## 
## 
## Eric Lecoutre: I don't want to die being idiot...
## Peter Dalgaard: With age, most of us come to realise that that is the only
## possible outcome.
##    -- Eric Lecoutre and Peter Dalgaard
##       R-help (October 2004)
## 
## 
## Morten Welinder: The result changes 70000 orders of magnitude right here. Ugh.
## This is affecting some erlang calculations of mine pretty badly.
## Peter Dalgaard: Make that 30400 orders of magnitude (natural logs y'know)... On
## something that's about 300000 orders of magnitude below 1, mind you! What the
## devil are you calculating? The probability that a random configuration of atoms
## would make up the known universe?
##    -- Morten Welinder and Peter Dalgaard (reporting a pgamma discontinuity)
##       R-devel (October 2004)
## 
## 
## Firstly, don't call your matrix 'matrix'. Would you call your dog 'dog'?
## Anyway, it might clash with the function 'matrix'.
##    -- Barry Rowlingson
##       R-help (October 2004)
## 
## 
## R is the lingua franca of statistical research. Work in all other languages
## should be discouraged.
##    -- Jan de Leeuw (as quoted by Matt Pocernich on R-help)
##       JSM 2003, San Francisco (August 2003)
## 
## 
## Amer Siddique: looking for some thoughts on incorporating R functionality to
## create histograms of data stored in an informix db. im gonna write the app in
## .Net and will use a managed provider to access the data. what R libs might I
## have to package in the assemblies? (sorry my Q is general as Ive only just
## looked at wanting this yet)
## Brian D. Ripley: Could you translate the question into English? My guess is
## that you are writing in a lower-cased version of some Microsoft internal
## language, but you don't even mention that this is for Windows (if it is).
##    -- Amer Siddique and Brian D. Ripley (on using R in .NET apps)
##       R-help (November 2004)
## 
## 
## Tony Plate: It is the prerogative of the members of R-core to decide whether to
## spend their time on improving documentation or improving the functionality (or
## merely having a life).
## Patrick Burns: I thought we had agreed that members of R-core are not allowed
## lives.
##    -- Tony Plate and Patrick Burns (in a thread about how to improve the
##       documentation of R functions)
##       R-devel (November 2004)
## 
## 
## Obviously I have not tested on an Alphabc...tuvwxyz system but it should work
## there.
##    -- Brian D. Ripley (after fixing a problem discovered when compiling R 2.0.0
##       on Linux Alpha)
##       R-devel (October 2004)
## 
## 
## Lars Strand: Will R run under Windows Pocket PC?
## Brian D. Ripley: We don't know! There are no binary versions of R for that
## platform, but perhaps you could find a suitable compiler and manage to build
## the sources.
## Outside pure mathematics it is usually very hard to establish that something
## cannot be done (and it can be very hard in pure mathematics, too).
##    -- Lars Strand and Brian D. Ripley
##       R-help (November 2004)
## 
## 
## It [the effort of learning how to use R] is the price paid, just as the dollars
## or euros for a commercial package would be. For that price, I've learnt a great
## deal - and not only about R. And I shall remember it when I next have to find a
## heavyweight solution for a big problem presented by a small charitable client
## with an invisible budget. It's a huge, awe-inspiring package - easier to
## perceive as such because the power is not hidden beneath a cosmetic veneer.
##    -- Felix Grant (in an article about free statistics software)
##       Scientific Computing World (November 2004)
## 
## 
## There are companies whose yearly license fees to SAS total millions of dollars.
## Then those companies hire armies of SAS programmers to program an archaic macro
## language using old statistical methods to produce ugly tables and the worst
## graphics in the statistical software world.
##    -- Frank Harrell
##       R-help (November 2004)
## 
## 
## I believe functions should do what they say on the box (and the help page), and
## not what some user hopes they might do by mind-reading.
##    -- Brian D. Ripley
##       R-help (November 2004)
## 
## 
## Thomas Lumley: The algorithm in glm.fit, while not perfect, is a little smarter
## than a simple IRLS. It uses step-halving to back away from the edge, and when
## the parameter space is convex it has a reasonable chance of creeping along the
## boundary to the true MLE.
## Peter Dalgaard: Hmm. That wasn't my experience. I had a situation where there
## was like a (virtual) maximum outside the boundary, and the algorithm would
## basically stay on the path to that peak, banging its little head into the same
## point of the wall repeatedly, so to speak.
##    -- Thomas Lumley and Peter Dalgaard (about problems with constrained
##       optimzation in GLMs and "the little optimizer that couldn't")
##       R-help (November 2004)
## 
## 
## Getting flamed for asking dumb questions on a public mailing list is all part
## of growing up and being a man/woman.
##    -- Michael Watson (in a discussion on whether answers on R-help should be
##       more polite)
##       R-help (December 2004)
## 
## 
## You may have not been long enough on this list to see that some of the old-time
## gurus have reached a demigod like status. Demigods have all rights to be 'rude'
## (that's almost a definition of a demi-deity).
##    -- Jari Oksanen (in a discussion on whether answers on R-help should be more
##       polite)
##       R-help (December 2004)
## 
## 
## Suggesting a Usenet newsgroup is a generation thing. I think some of the
## younger users would prefer a Wiki or a Forum (these are words I've seen, but I
## wouldn't visit places like this, talking about my g-g-generation).
##    -- Jari Oksanen
##       R-help (December 2004)
## 
## 
## Our great-great grandchildren as yet unborn may read some of the stupid
## questions and/or answers that I and perhaps others give from time to time. I'd
## rather get flamed for saying something stupid in public on this list than to
## continue to provide substandard service to the people with whom I work because
## I perpetrated the same mistake in an environment in which no one questioned so
## effectively my errors.
##    -- Spencer Graves (in a discussion on whether answers on R-help should be
##       more polite)
##       R-help (December 2004)
## 
## 
## Robin Hankin: I'd say that without a tool like R you cannot learn statistics.
## David Whiting: I believe Fisher and a few others managed to get by without it.
## Peter Dalgaard: But think how far they could have got with R!
##    -- Robin Hankin, David Whiting, and Peter Dalgaard (on teaching/learning
##       statistics with R)
##       R-help (December 2004)
## 
## 
## If you don't go with R now, you will someday.
##    -- David Kane (on whether to use R or S-PLUS)
##       R-SIG-Finance (November 2004)
## 
## 
## Patrick Burns: In the old days with S-PLUS, the rule of thumb was that you
## needed 10 times as much memory as your dataset. [...] R (and current versions
## of S-PLUS) are more frugal than S-PLUS was back then.
## Ajay Shah: Hmm, so it'd be interesting to apply current prices of RAM and
## current prices of R, to work out the break-even point at which it's better to
## buy SAS! :-) Without making any calculations, I can't see how SAS can compete
## with the price of 4G of RAM.
##    -- Patrick Burns and Ajay Shah (on whether to use R, S-PLUS or SAS)
##       R-SIG-Finance (December 2004)
## 
## 
## I have never heard anyone (knowledgable or otherwise) claim that, in the
## absence of transition costs, SAS is better than R for equity modeling. If you
## come across any such claim, I would be happy to refute it.
##    -- David Kane
##       R-SIG-Finance (December 2004)
## 
## 
## Will Frank Harrell or someone else please explain to me a real application in
## which this is not fast enough?
##    -- Brian D. Ripley (announcing improved versions of read.table and
##       write.table)
##       R-devel (December 2004)
## 
## 
## Achim Zeileis: Other visualizations of departures from indepence include
## mosaicplots with shading or association plots. See mosaicplot() and assocplot()
## and the package vcd.
## Peter Dalgaard: I see that you've taken initive to abbriate statiscal
## concepts...
##    -- Achim Zeileis and Peter Dalgaard
##       R-help (January 2005)
## 
## 
## Also, that's what Martin Maechler told me and he wrote the code so I trust him
## on that. I figure that if you have to trust someone to be meticulous and
## precise then a German-speaking Swiss is a good choice.
##    -- Douglas Bates
##       R-help (January 2005)
## 
## 
## Paul Gilbert: [code comparing speed of apply(z,2,sum) vs. rep(1,10000)%*%z)]
## which seemed completely contrary to all my childhood teachings.
## Douglas Bates: Must have had an interesting childhood if you spent it learning
## about the speeds of various matrix multiplication techniques.
## Paul Gilbert: [...] why is apply so slow?
## Brian Ripley: 'so slow' sic: what are you going to do in the 7ms you saved?
##    -- Paul Gilbert, Douglas Bates, and Brian D. Ripley (discussing 'the
##       incredible lightness of crossprod')
##       R-devel (January 2005)
## 
## 
## Well, if I had written that, it would have been on purpose ...
##    -- Kurt Hornik (after Torsten Hothorn reported the 'typo' automa*g*ically in
##       ?NotYetImplemented)
##       R-devel (January 2005)
## 
## 
## I'm not sure I'd trust any computer recommendation from 1976, no matter how
## famous the authors are.
##    -- Peter Dalgaard (after Samuel Edward Kemp cited a recommendation about
##       nonlinear least squares computer programs from 'Box-Jenkins, 1976')
##       R-help (January 2005)
## 
## 
## Marc R. Feldesman: I'm trying to figure out what could have possibly changed
## [...] that make Brian Ripley's posts (in particular) not show up on my mail
## server [...]
## Peter Dalgaard: Did you remember to turn off the Oxford-sarcasm filter?
##    -- Marc R. Feldesman and Peter Dalgaard (Marc R. Feldesman having some
##       e-mail problems)
##       R-help (January 2005)
## 
## 
## Mingzhai Sun: When you use it [R], since it is written by so many authors, how
## do you know that the results are trustable?
## Bill Venables: The R engine [...] is pretty well uniformly excellent code but
## you have to take my word for that. Actually, you don't. The whole engine is
## open source so, if you wish, you can check every line of it. If people were out
## to push dodgy software, this is not the way they'd go about it.
##    -- Mingzhai Sun and Bill Venables
##       R-help (January 2004)
## 
## 
## Jesus and the rest of the R-help community: Thanks for your help.
##    -- Mike Saunders (after Jesus Frias answered his question about split-split
##       plots)
##       R-help (February 2005)
## 
## 
## Pavel Khomski: How can I specify the random part in the GLMM-call (of the lme4
## library) for compound matrices just in the the same way as they defined in the
## lme-Call (of the nlme library).
## Martin Maechler: ''twice in such a short paragraph -- yikes !!'' ... I'm
## getting convulsive...
##    -- Pavel Khomski and Martin Maechler
##       R-help (February 2005)
## 
## 
## A sufficiently trained statistician can read the vagaries of a Q-Q plot like a
## shaman can read a chicken's entrails, with a similar recourse to scientific
## principles. Interpreting Q-Q plots is more a visceral than an intellectual
## exercise. The uninitiated are often mystified by the process. Experience is the
## key here.
##    -- Department of Mathematics and Statistics, Murdoch University
##       StatsNotes
## 
## 
## If the answer is parse() you should usually rethink the question.
##    -- Thomas Lumley
##       R-help (February 2005)
## 
## 
## Just now I had an apparently insurmountable problem that's been bugging me for
## days, but phrasing my question in a form suitable for the R-help list enabled
## me to solve my own problem in two minutes flat.
## Thanks everyone.
##    -- Robin Hankin
##       R-help (March 2005)
## 
## 
## Actually, I see it as part of my job to inflict R on people who are perfectly
## happy to have never heard of it. Happiness doesn't equal proficient and
## efficient. In some cases the proficiency of a person serves a greater good than
## their momentary happiness.
##    -- Patrick Burns
##       R-help (April 2005)
## 
## 
## Evelyn Hall: I would like to know how (if) I can extract some of the
## information from the summary of my nlme.
## Simon Blomberg: This is R. There is no if. Only how.
##    -- Evelyn Hall and Simon 'Yoda' Blomberg
##       R-help (April 2005)
## 
## 
## Tradition among experienced S programmers has always been that loops (typically
## 'for' loops) are intrinsically inefficient: expressing computations without
## loops has provided a measure of entry into the inner circle of S programming.
##    -- John Chambers
##       Programming With Data, p. 173 (1998)
## 
## 
## You can't expect statistical procedures to rescue you from poor data.
##    -- Berton Gunter (on dealing with missing values in a cluster analysis)
##       R-help (April 2005)
## 
## 
## I think we can reject the null hypothesis of "Dirk can type" at all convential
## significance levels.
##    -- Dirk Eddelbuettel (after several users including himself had misspelled
##       Ubuntu as Umbutu or Ubunto)
##       R-help (April 2005)
## 
## 
## The problem, as always, is what the heck does one mean by 'outlier' in these
## contexts. Seems to be like pornography -- "I know it when I see it."
##    -- Berton Gunter (quoting Justice Potter Stewart in a discussion about tests
##       for outliers)
##       R-help (April 2005)
## 
## 
## [...] (damn... my shift key just broke as well...)
## bill venables.
##    -- Bill Venables (answering a message that lacked upper case letters)
##       R-help (April 2005)
## 
## 
## It is a mistake to assume that everything is known about even the common
## approximations used in statistical practice, and this area is still opening up.
##    -- Bill Venables (in a discussion about how fitted generalized linear mixed
##       models can be compared)
##       R-help (April 2005)
## 
## 
## We are teaching stuff like mathematics, statistics, statistical programming and
## so on, but we cannot teach each possible software product - we also don't teach
## how to analyze each possible dataset that might find its way to the
## statistician's desk, we just use the iris data and Anscombe's quartet. ;-)
##    -- Uwe Ligges (in a thread about whether universities have to teach students
##       how to use commercial software)
##       R-help (April 2005)
## 
## 
## Usually (but not always) doing tests of normality reflect a lack of
## understanding of the power of rank tests, and an assumption of high power for
## the tests (qq plots don't always help with that because of their subjectivity).
## When possible it's good to choose a robust method. Also, doing pre-testing for
## normality can affect the type I error of the overall analysis.
##    -- Frank Harrell
##       R-help (April 2005)
## 
## 
## As for elegance, R is refined, tasteful, and beautiful. When I grow up, I want
## to marry R.
##    -- Andy Bunn (in a discussion about whether R is simple, powerful and
##       elegant)
##       R-help (May 2005)
## 
## 
## Frank Harrell: Here is an easy approach that will yield results only slightly
## less valid than one actually using the response variable:
##   x <- data.frame(x1, x2, x3, x4, ..., other potential predictors)
##   x[ , sample(ncol(x))]
## Andy Liaw: Hmm... Shouldn't that be something like:
##   x[, sample(ncol(x), ceiling(ncol(x) * runif(1)))]
##    -- Frank Harrell and Andy Liaw (about alternative strategies for stepwise
##       regression and 'random parsimony')
##       R-help (May 2005)
## 
## 
## Rene M. Raupp: Does anybody know any work comparing R with other (charged)
## statistical software (like Minitab, SPSS, SAS)? [...] I have to show it's as
## good as the others.
## Kjetil Brinchmann Halvorsen: Sorry. That will be difficult. Couldn't it do to
## prove it is better?
##    -- Rene M. Raupp and Kjetil Brinchmann Halvorsen
##       R-help (May 2005)
## 
## 
## Sean Davis: It got me going quickly with S4 methods, which it seems to me are
## the way to go in most cases.
## Rolf Turner: If you want to simultaneously handcuff yourself, strap yourself
## into a strait jacket, and tie yourself in knots, and moreover write code which
## is incomprehensible to the human mind, then S4 methods are indeed the way to go.
##    -- Sean Davis and Rolf Turner (expressing different views about the benefits
##       of S4 classes)
##       R-help (May 2005)
## 
## 
## I wish to perform brain surgery this afternoon at 4pm and don't know where to
## start. My background is the history of great statistician sports legends but I
## am willing to learn. I know there are courses and numerous books on brain
## surgery but I don't have the time for those. Please direct me to the
## appropriate HowTos, and be on standby for solving any problem I may encounter
## while in the operating room. Some of you might ask for specifics of the case,
## but that would require my following the posting guide and spending even more
## time than I am already taking to write this note.
##    -- I. Ben Fooled (aka Frank Harrell)
##       R-help (April 1, 2005)
## 
## 
## You're overlooking something like line 800 of the documentation for xyplot.
## [...] It's probably in the R-FAQ as well, since my original feeling was that
## this behaviour was chosen in order to confuse people and see how many people
## read the FAQ... :)
##    -- Barry Rowlingson (about the fact that lattice objects have to be
##       print()ed)
##       R-help (May 2005)
## 
## 
## Rolf Turner: In the middle of a Saturday morning (in my Time Zone!) I send out
## a plea for help, and in just over 20 minutes my problem is solved!
## I don't think you get service like that anywhere else. This R-help list is
## BLOODY AMAZING!
## Spencer Graves: 'The sun never sets on the (former) British Empire.' Today, it
## never sets on R-Help.
##    -- Rolf Turner and Spencer Graves
##       R-help (May 2005)
## 
## 
## As a general rule (subject to numerous exceptions, caveats, etc.):
## 1) it is programming and debugging time that most impacts 'overall' program
## execution time;
## 2) this is most strongly impacted by code readability and size (the smaller the
## better);
## 3) both of which are enhanced by modular construction and reuseability, which
## argues for avoiding inline code and using separate functions.
## These days, I would argue that most of the time it is program clarity and
## correctness (they are related) that is the important issue, not execution speed.
##    -- Berton Gunter (in a discussion about parsing speed)
##       R-help (May 2005)
## 
## 
## I suggest that we redefine "+" to be a legal character in the name of an
## identifier. Sure, if you used to have "a+b" you'll have to rewrite it as
## "+(a,b)", but think of the benefits!
## You+can+write+out+long+sentences+and+use+them+as+variable+names!
##    -- Duncan Murdoch (in a discussion whether to use underscore, dot, or mixed
##       case in variable names)
##       R-help (May 2005)
## 
## 
## Oops. Was that what I was expecting?
##    -- Louise Matthews (during the practicals of 'Statistical Practice in
##       Epidemiology Using R')
##       SPE Course, Tartu (May 2005)
## 
## 
## Well, 'do as we say, not do as we do' applies, I think.
##    -- Brian D. Ripley (answering a report on a small inconsistency in 'Writing
##       R Extensions' by David Firth)
##       R-devel (June 2005)
## 
## 
## See pages 45 and 46 of V&R's S PROGRAMMING for a definitive discussion of how
## to do this.
## (Confession: I treat anything that V&R say as definitive).
##    -- Berton Gunter (in reply to the question how to parse a '...' function
##       argument)
##       R-help (June 2005)
## 
## 
## A big computer, a complex algorithm and a long time does not equal science.
##    -- Robert Gentleman
##       SSC 2003, Halifax (June 2003)
## 
## 
## Uwe Ligges: I just told nonsense, stepclass() does not make sense with
## randomForest(), obviously ... (wonder why nobody shouted?).
## Douglas Bates: Oh, we're just so used to you talking nonsense that we don't
## bother to point it out any more :-)
##    -- Uwe Ligges and Douglas Bates
##       R-help (July 2005)
## 
## 
## The fact that some people murder doesn't mean we should copy them. And
## murdering data, though not as serious, should also be avoided.
##    -- Frank E. Harrell (answering a question on categorization of continuous
##       variables in survival modelling)
##       R-help (July 2005)
## 
## 
## The learning curve is steep - but then like many people, I'd like to be able to
## do sophisticated modelling with deep understanding and no effort :-)
##    -- Sean O'Riordain (in a thread about the helpfulness of documentation)
##       R-help (July 2005)
## 
## 
## Excellent, Berwin!
## Specifically the part about why the bug never really triggers wrong behavior.
## Your mail should be saved in 'The Annals of R' ...
##    -- Martin Maechler (after Berwin A Turlach reported a buglet in
##       'spline_eval')
##       R-devel (July 2005)
## 
## 
## You say yourself it wasn't reproducible. So it could have been anything that
## "crashed" your R, cosmic radiation, a bolt of lightning reversing a bit in your
## computer memory, ... :-)
##    -- Martin Maechler (replying to a bug report)
##       R-devel (July 2005)
## 
## 
## It was simple, but you know, it's always simple when you've done it.
##    -- Simone Gabbriellini (after solving a problem with a trick suggested on
##       the list)
##       R-help (August 2005)
## 
## 
## Trevor Hastie: It would be great to have a 'clickable' digest, where the topics
## list is a set of pointers, and clicking on a topic takes you to that entry.
## Jeff Gentry: I sincerely hope you don't mean that you wish HTML enabled content
## in the actual emails coming through the mailing list. People who send HTML in
## email text should be taken out back and shot :)
## Peter Dalgaard: Pretty much what my spam filter does... Well, it kills the
## message, not the sender, although the latter would be a more permanent solution.
##    -- Trevor Hastie, Jeff Gentry, and Peter Dalgaard
##       R-help (August 2005)
## 
## 
## If you aren't sure of the relative priority of two operators, use parentheses.
## 1:(nr-1) would work regardless of whether : or - had higher priority. Or, in
## extreme cases, read the documentation.
##    -- Duncan Murdoch
##       R-help (August 2005)
## 
## 
## It is wrong to blame ANY tool for our own shortcomings!
##    -- Robert Baer (about the perils of using PowerPoint)
##       R-help (September 2005)
## 
## 
## Jason Horn: Can anyone think of a way to create a pretty() sequence that
## excludes zero? Or a way to remove the zero from a sequence after using pretty()?
## Peter Dalgaard: The former is rather hard because zero is generally considered
## just about the prettiest number around...
##    -- Jason Horn and Peter Dalgaard
##       R-help (October 2005)
## 
## 
## Does anyone know someone working in this area in France who is not a JARP (Just
## Another R Person)?
##    -- Jan de Leeuw
##       JSS mailing list (October 2005)
## 
## 
## Contrary to popular belief the speed of R's interpreter is rarely the limiting
## factor to R's speed. People treating R like C is typically the limiting factor.
## You have vector operations, USE THEM.
##    -- Byron Ellis
##       R-help (October 2005)
## 
## 
## Mathematical conventions are just that, conventions. They differ by field of
## mathematics. Don't ask us why matrix rows are numbered down but graphs are
## numbered up the y axis, nor why x comes before y but row before column. But the
## matrix layout has always seemed illogical to me.
##    -- Brian D. Ripley (answering a question why the layouts of print(x) and
##       image(x) are different)
##       R-help (August 2004)
## 
## 
## Byron Ellis: If we wanted to be truly radical we'd just accept that graphics
## devices and event loops are just special cases of the connection and merge the
## whole thing, thus more-or-less reinventing CLIM. :-)
## Anthony Rossini: Eventually, all programming languages grow up and become Lisp.
## (progress, progress, and more joyful progress on CLS).
## Byron Ellis: Untrue! They may also become Smalltalk :-)
##    -- Byron Ellis and Anthony Rossini
##       R-devel (December 2005)
## 
## 
## And the fear of getting Ripleyed on the mailing list also makes me think, read,
## and improve before submitting half baked questions to the list.
##    -- Eric Kort
##       R-help (January 2006)
## 
## 
## Ole F. Christensen: Brain, Thank you very much for your help.
## Peter Dalgaard: You seem to be using call-by-value semantics rather than
## call-by-name...
##    -- Ole F. Christensen and Peter Dalgaard (thanking *Brian* D. Ripley for
##       help)
##       R-devel (January 2006)
## 
## 
## Thomas Lumley: The precedence of ^ is higher than that of unary minus. It may
## be surprising, [...]
## Herve Pages: No, it's not surprising. At least to me... In the country where I
## grew up, I've been teached that -x^2 means -(x^2) not (-x)^2.
##    -- Thomas Lumley and Herve Pages (both explaining that operator precedence
##       is working perfectly well)
##       R-devel (January 2006)
## 
## 
## The wiki spamming is a serious problem, [...]. What does a logfile with IPs
## help? The spammers are strangers selling viagra. I don't want to find them :-)
##    -- Detlef Steuer (in a discussion on R wikis)
##       R-help (January 2006)
## 
## 
## Any doubts about R's big-league status should be put to rest, now that we have
## a Sudoku Puzzle Solver. Take that, SAS!
##    -- David Brahm (announcing the sudoku package)
##       R-packages (January 2006)
## 
## 
## This list is impressive. People are knowledgable, opinionated, ready to help
## and to flame you for asking elementary question or asking how to use type III
## SSQ. So, speak softly and carry a beagle. Seriously, sometimes it would be
## quicker just to give an answer, than to flame a poor soul.
##    -- Milos Zarkovic
##       R-help (January 2006)
## 
## 
## Lastly, by what we could call the 'Liaw-Baron principle', every question that
## can be asked has in fact already been asked.
##    -- Dirk Eddelbuettel (citing Andy Liaw's and Jonathan Baron's opinion on
##       unique questions on R-help)
##       R-help (January 2006)
## 
## 
## It's not a question of trying variations, rather of following instructions.
##    -- Brian D. Ripley (about using 'Writing R Extensions')
##       R-help (January 2006)
## 
## 
## Michael Dewey: When I started I found the free documents useful but I made most
## progress when I bought MASS. I do realise that liking books is a bit last
## millennium.
## Thomas Lumley: Very late last millenium, though. "When I were young[er] we
## didn't have all these fancy yellow books."
##    -- Michael Dewey and Thomas Lumley (about different kinds of documentation
##       for R)
##       R-help (January 2006)
## 
## 
## Please do think hard before you tell other people what they 'should' do for you.
##    -- Brian D. Ripley
##       R-devel (January 2006)
## 
## 
## It might surprise many R-help posters, but R has manuals as well...
##    -- Uwe Ligges
##       R-help (January 2006)
## 
## 
## Tom Backer Johnsen: I have just started looking at R, and are getting more and
## more irritated at myself for not having done that before. However, one of the
## things I have not found in the documentation is some way of preparing output
## from R for convenient formatting into something like MS Word.
## Barry Rowlingson: Well whatever you do, don't start looking at LaTeX, because
## that will get you even more irritated at yourself for not having done it before.
##    -- Tom Backer Johnsen and Barry Rowlingson
##       R-help (February 2006)
## 
## 
## Michael Watson: Hopefully this one isn't in the manual or I am about to get
## shot :-S
## Peter Dalgaard: *Kapow*...
##    -- Michael Watson and Peter Dalgaard (question on axis())
##       R-help (February 2006)
## 
## 
## My solution when I run into mysteries like this is to put 'browser()' in the
## function just before or after the line of interest. The magnitude and direction
## of my stupidity usually become clear quickly.
##    -- Patrick Burns
##       R-help (February 2006)
## 
## 
## Similarly to Brian, I'm much more reluctant to help people who don't exist --
## who knows, maybe you're a computer program who has just passed the Turing test
## :-)
##    -- Martin Maechler (about postings with anonymous e-mail addresses and
##       without real names)
##       R-help (March 2006)
## 
## 
## There's probably some examples, but there are some examples of people using
##   solve(t(X) %*% W %*% X) %*% W %*% Y
## to compute regression coefficients, too.
##    -- Thomas Lumley (discussing usefulness of evaluation order in lapply)
##       R-help (March 2006)
## 
## 
## (3 times...rrrrgh...) and why do you think the mailing list is called
## R-*packages* ???????????
## Please do
##   for(i in 1:20) cat("It's a package!\n")
##    -- Martin Maechler (after a newly released *package* has been called
##       *library* three times in its announcement on R-packages)
##       R-help (March 2006)
## 
## 
## If we do that, I predict that the total amount of r-help time wasted on it will
## exceed the CPU time saved by orders of magnitude.
##    -- Duncan Murdoch (after a specialized function speeding up detection of NAs
##       was suggested for base R)
##       R-devel (April 2006)
## 
## 
## This has been discussed before in this list, and Ripley said "no, no!". I do it
## all the time, but only in secrecy.
##    -- Jari Oksanen (about replacing zero distances with tiny values for
##       isoMDS())
##       R-help (April 2006)
## 
## 
## 2-D pie charts are terrible. That makes 3-D pie charts terrible to the 3/2
## power.
##    -- Frank Harrell
##       R-help (April 2006)
## 
## 
## Yourexamplesmightactuallybereadableifyouusedyourspacebar.
##    -- Brian D. Ripley (answering a question containing code without any spaces)
##       R-help (May 2006)
## 
## 
## It seems that you are facing a very serious fortune(122) problem.
##    -- Jean R. Lobry (in reply to a user that quickly needed help for his PhD
##       thesis, without properly checking the documentation)
##       R-help (May 2006)
## 
## 
## Duncan Murdoch: Others need to run under ESS.
## Francois Pinard: While this is a good things for Emacs lovers, the requirement
## is rather unwelcome for pagans! :-)
##    -- Duncan Murdoch and Francois Pinard (on the availability of command
##       completion)
##       R-help (May 2006)
## 
## 
## When talking about user friendliness of computer software I like the analogy of
## cars vs. busses: [...]
## Using this analogy programs like SPSS are busses, easy to use for the standard
## things, but very frustrating if you want to do something that is not already
## preprogrammed.
## R is a 4-wheel drive SUV (though environmentally friendly) with a bike on the
## back, a kayak on top, good walking and running shoes in the passenger seat, and
## mountain climbing and spelunking gear in the back. R can take you anywhere you
## want to go if you take time to learn how to use the equipment, but that is
## going to take longer than learning where the bus stops are in SPSS.
##    -- Greg Snow
##       R-help (May 2006)
## 
## 
## Oh, gosh, this is getting whimsical.
## R, I surmise, is a mutable tool that enables the user to do what they want.
## Knapped flint for the Stone Age statistical fool, plasma arc for the expert and
## learned savant.
## R is a friend to all manner of men. The perfect companion, the servant
## complete. It gently informeth the clueless and then, reveals the essence unto
## the esthete.
##    -- Jim Lemon (in a discussion about the usability of R)
##       R-help (May 2006)
## 
## 
## Statistical computing is not easy, so how could R be? Who has ever claimed it
## is? Any package that makes statistical computing appear to be easy is probably
## giving you wrong answers half the time, or is extremely limited in scope.
##    -- Duncan Murdoch (in a discussion about the usability of R)
##       R-help (May 2006)
## 
## 
## Jarek Tuszynski: At one company [...] I had to run all the licenses of all the
## software [...] through the legal department. When they read the GNU Public
## License (GPL) their only comment was: We have no idea what that license means.
## Do not touch any software using it.
## Rolf Turner: This is typical of lawyers' minds. If something is clear,
## rational, lucid, straightforward, unambigous, means what it says ... they can't
## understand it.
##    -- Jarek Tuszynski and Rolf Turner (in a discussion about the GPL and
##       'buying R')
##       R-help (May 2006)
## 
## 
## It is unusual for the actual data not to be available in real problems.
##    -- Brian D. Ripley (in reply to a question how to fit a distribution if not
##       the data but only their histogram is available)
##       R-help (June 2006)
## 
## 
## Warning: Although abusing R was not proved to be addictive, it should be noted
## that it often leads to harder stuff.
##    -- Ivan Mizera (self-confessed abuseR)
##       useR! 2006, Vienna (June 2006)
## 
## 
## I wish <<- had never been invented, as it makes an esoteric and dangerous
## feature of the language *seem* normal and reasonable. If you want to dumb down
## R/S into a macro language, this is the operator for you.
##    -- Bill Venables
##       R-help (July 2001)
## 
## 
## John Kane: I have 120 columns in a data.frame. I have one value in a column
## named "blaw" that I want to change. How do I find the coordinates?
## Roger Koenker: It is the well-known wicked which problem: If you had
## (grammatically incorrectly) thought "... which I want to change" then you might
## have been led to type (in another window):
##   ?which
## and you would have seen the light. Maybe that() should be an alias for which()?
##    -- John Kane and Roger Koenker
##       R-help (August 2006)
## 
## 
## Gregor Gorjanc: But imagine how hard would it be to have two separate modes ...
## argh, probably a mess^2 or have I missed something obvious.
## Martin Maechler: Yes, exactly: "Mess ^ 2" -- and if you allow both 'drop' and
## 'na.rm' options, it's "Mess ^ 3" -- not something anyone really wants!
##    -- Gregor Gorjanc and Martin Maechler (in a discussion whether 'drop' should
##       be set in options(), possibly depending on different user vs. programmer
##       modes)
##       R-devel (September 2006)
## 
## 
## You need to prepare your students for the future, not the past.
##    -- Spencer Graves (making a case for using R in academia)
##       R-help (November 2006)
## 
## 
## Have you ever wanted to write a book, but not known where to start? Now is a
## very good time to jump in, because there is currently a very simple recipe for
## success: just put R in the title and you will have to beat the publishers off
## with a stick!
##    -- Paul Murrell
##       ASA Statistical Computing & Graphics Newsletter 17(2) (November 2006)
## 
## 
## This is at best a matter of opinion, and credentials do matter for opinions.
##    -- Brian D. Ripley (in response to the claim that only 1.79e3 is printed
##       with three significant digits whereas 1790 is not)
##       R-help (December 2006)
## 
## 
## If you want to do an integer divide, you should only use integers to divide
## with.
##    -- Thierry Onkelinx ((not entirely correctly) in an answer to "understanding
##       integer divide (%/%)")
##       R-help (January 2007)
## 
## 
## Personally I have never regretted trying not to underestimate my own future
## stupidity.
##    -- Greg Snow (explaining why eval(parse(...)) is often suboptimal, answering
##       a question triggered by the infamous fortune(106))
##       R-help (January 2007)
## 
## 
## Brian Ripley: Where did you tell it [...]? (Nowhere: R is lacking a mind_read()
## function!)
## Peter Dalgaard: Please stop complaining about missing features. Patches will be
## considered.
## 
## Oh, it's you, Brian. Never mind then. You'll get to it, I'm sure. ;-)
##    -- Brian Ripley and Peter Dalgaard (answering why abline(lm(x~y)) would not
##       work)
##       R-help (January 2007)
## 
## 
## Just [use] MinGW like R [does], following the guides to the letter gets you
## there like marked stones across a marsh. Leaving the path usually gets you at
## best neck deep in the mire, alternatively just bubbles.
##    -- Roger Bivand (explaining what development environment should be used to
##       develop C on Windows (for R))
##       R-devel (January 2007)
## 
## 
## You must realize that R is written by experts in statistics and statistical
## computing who, despite popular opinion, do not believe that everything in SAS
## and SPSS is worth copying. Some things done in such packages, which trace their
## roots back to the days of punched cards and magnetic tape when fitting a single
## linear model may take several days because your first 5 attempts failed due to
## syntax errors in the JCL or the SAS code, still reflect the approach of "give
## me every possible statistic that could be calculated from this model, whether
## or not it makes sense". The approach taken in R is different. The underlying
## assumption is that the useR is thinking about the analysis while doing it.
##    -- Douglas Bates (in reply to the suggestion to include type III sums of
##       squares and lsmeans in base R to make it more similar to SAS or SPSS)
##       R-help (March 2007)
## 
## 
## I don't like to see the use of c() for its side effects. In this case Marc's
## as.vector seems to me to be self-explanatory, and that is a virtue in
## programming that is too often undervalued.
##    -- Brian D. Ripley (on how to convert a matrix into a vector)
##       R-help (March 2007)
## 
## 
## I cannot remember if I have been using 14 or 14, I think it was 14 and I am not
## near the machine to check.
##    -- John Kane (confused about his SPSS version)
##       R-help (April 2007)
## 
## 
## Tony Plate: There looks to be a typo in the R-exts manual: [...]
## Peter Dalgaard: 'svn blame' tells me that this was Brian's addition in
## rev.35362 [...]
## Brian D. Ripley: I prefer 'svn praise' myself.
## Peter Dalgaard: Or 'svn annotate'. I think it depends on what I'm looking for,
## plus the risk that the author (perpetrator, contributor) might be me...
##    -- Tony Plate, Peter Dalgaard, and Brian D. Ripley (about a typo in the
##       documentation)
##       R-devel (April 2007)
## 
## 
## Corinna Schmitt: How can I divide the number 0.285 with 2. I need a function.
##   Result: 0.285 / 2 = 0.1425
## Gabor Csardi: Well, i think
##   half.of.0.285 <- function() {
##     0.1425
##   }
## would do the trick.
##    -- Corinna Schmitt and Gabor Csardi
##       R-help (April 2007)
## 
## 
## Perhaps one is the real forge and the other is a forgery? Or a forge-R-y? I'll
## get my coat...
##    -- Barry Rowlingson (on the question whether http://www.RForge.net/ or
##       http://R-Forge.R-project.org/ is the official forge server)
##       R-help (April 2007)
## 
## 
## It is all too easy for statistical thinking to be swamped by programming tasks.
##    -- Brian D. Ripley
##       DSC 2001, Wien (March 2001)
## 
## 
## R is being used for different purposes than the original interactive EDA
## environment.
##    -- Duncan Temple Lang
##       DSC 2007, Auckland (February 2007)
## 
## 
## RAM is cheap and thinking hurts.
##    -- Uwe Ligges (about memory requirements in R)
##       R-help (June 2007)
## 
## 
## All this becomes even more glaring if you take the unusal step of plotting the
## data.
##    -- Bill Venables (interpreting the results of an ANOVA analysis)
##       R-help (July 2007)
## 
## 
## What sort of editor would overlook this clear and demonstrable message leaping
## out from the data in favour of some arcane argument about "types of sums of
## squares"? Several answers come to mind: A power freak, a SAS afficianado, an
## idiot.
##    -- Bill Venables
##       R-help (July 2007)
## 
## 
## If you really want to assess uncertainty you need to take into account that the
## models are false and that several models may capture different aspects of the
## data and so be false in different ways.
##    -- Brian D. Ripley
##       R-help (July 2007)
## 
## 
## R will always be arcane to those who do not make a serious effort to learn it.
## It is **not** meant to be intuitive and easy for casual users to just plunge
## into. It is far too complex and powerful for that. But the rewards are great
## for serious data analysts who put in the effort.
##    -- Berton Gunter
##       R-help (August 2007)
## 
## 
## If anything, there should be a Law: Thou Shalt Not Even Think Of Producing A
## Graph That Looks Like Anything From A Spreadsheet.
##    -- Ted Harding (in a discussion about producing graphics)
##       R-help (August 2007)
## 
## 
## Friends don't let friends use Excel for statistics!
##    -- Jonathan D. Cryer (about problems with using Microsoft Excel for
##       statistics)
##       JSM 2001, Atlanta (August 2001)
## 
## 
## Anand Patil: Can this be fixed without requiring the user to do anything?
## Brian D. Ripley: The fix requires 'the user' to read the documentation.
##    -- Anand Patil and Brian D. Ripley (about some compiler problems under
##       Windows Vista)
##       R-help (September 2007)
## 
## 
## Bad English is the language of science.
##    -- Jan de Leeuw
##       JSS mailing list (February 2005)
## 
## 
## 3-D bar plots are an abomination. Just because Excel can do them doesn't mean
## you should.
## (Dismount pulpit).
##    -- Berton Gunter
##       R-help (October 2007)
## 
## 
## Actually the documentation of sunflowerplot is wrong in botanical sense.
## Sunflowers have composite flowers in capitula, and the things called 'petals'
## in documentation are ligulate, sterile ray-florets (each with vestigial petals
## which are not easily visible in sunflower, but in some other species you may
## see three (occasionally two) teeth).
##    -- Jari Oksanen
##       R-help (December 2007)
## 
## 
## Apparently mathematical impossibility is not an impediment to parameter
## estimation in such cases.
##    -- Douglas Bates (about the abilities of SAS PROC MIXED)
##       R-help (December 2007)
## 
## 
## memory problems (not me. my pc!)
##    -- Sara Mouro (subject line for an R-help request)
##       R-help (January 2008)
## 
## 
## Gustaf Rydevik: The author also has some thought-provoking opinions on R being
## no-good and that you should write everything in C.
## Paul Gilbert: People used to say assembler, that's progress.
##    -- Gustaf Rydevik and Paul Gilbert (in a discussion about an 'R is slow'
##       blog post)
##       R-help (January 2008)
## 
## 
## Happy families are all alike; every unhappy family is unhappy in its own way.
## Leo Tolstoy
## 
## and every messy data is messy in its own way - it's easy to define the
## characteristics of a clean dataset (rows are observations, columns are
## variables, columns contain values of consistent types). If you start to look at
## real life data you'll see every way you can imagine data being messy (and many
## that you can't)!
##    -- Hadley Wickham (answering 'in what way messy data sets are messy')
##       R-help (January 2008)
## 
## 
## I strongly suggest you collaborate with a local statistician. I can think of no
## circumstance where multiple regression on "hundreds of thousands of variables"
## is anything more than a fancy random number generator.
##    -- Berton Gunter
##       R-help (February 2008)
## 
## 
## The "right" place to store information like this in Windows is in the registry,
## but we've been reluctant to use it, for various historical reasons. In fact, we
## were just told today that editing the registry is not allowed by software used
## by the US government. I think they'll be moving back to the abacus, if they
## want to stay with Windows.
##    -- Duncan Murdoch (about where to store some Windows GUI preferences)
##       R-devel (January 2008)
## 
## 
## I recently read the small print on the academic license our site has for SAS.
## You have to:
## 
##   1 inform SAS of any taught courses that use SAS,
##   2 inform SAS of any research projects using SAS,
##   3 allow SAS to refer to your institution as a SAS user,
##   4 allow SAS to review your taught courses,
##   5 ensure your courses are taught using qualified personnel,
##   6 give SAS your first-born male offspring.
## 
## I spoke to our site's licensing supremos and they say they've never heard of
## anyone complying with 1 or 2. Point 4 sounds like petty fiddling in our
## educational business, and point 5 left 'qualified' undefined. Point 6 doesn't
## bother me since I don't have kids.
##    -- Barry Rowlingson
##       R-help (February 2008)
## 
## 
## The "one size fits all" approach to data analysis - also known as "give me a
## quart and a half of statistics and just make sure that there is a p-value less
## than 5% somewhere in there" - doesn't fit well into the R system.
##    -- Douglas Bates
##       R-help (March 2008)
## 
## 
## In many cases a) often looks difficult, but on closer inspection turns out to
## be impossible.
##    -- Bill Venables (about calculating the negative log-likelihood from a
##       non-independence model)
##       R-help (April 2008)
## 
## 
## Basically R is reluctant to let you shoot yourself in the foot unless you are
## really determined to do so.
##    -- Bill Venables (about the warning hist() issues when being called with
##       unequal interval widths and freq=TRUE)
##       R-help (May 2008)
## 
## 
## Whoops! I read that e+24 as e-24, so scrub all that I said.
## (You'd think, that having seen Bill Venables make a similar error -- which he
## corrected in the follow-up posting to which I was replying -- I would've been
## more careful. Well, you'd think wrong. Actually it *was* e-24 when I posted;
## then the Gremlins got in and changed everything. :-) )
##    -- Rolf Turner (discussing that f.root from uniroot() is (not) almost zero)
##       R-help (May 2008)
## 
## 
## R may be the wrong tool for the job, but it's the wrong job.
##    -- Rolf Turner (about solving 100th degree polynomials)
##       R-help (May 2008)
## 
## 
## Kenn Konstabel: [...] There's more to this trend: SPSS and Statistica now
## advertise "R language support": [...]
## Charles C. Berry: If you can't beat R, join R.
## Marc Schwartz: "Resistance is futile. You will be assimilated." 
##    -- Kenn Konstabel, Charles C. Berry, and Marc Schwartz (about the RPro
##       release from REvolution)
##       R-help (August 2008)
## 
## 
## As for the question of the bug in `is', ... it depends what your definition of
## `is' is.
##    -- Douglas Bates (referring to Bill Clinton after a tortuous discussion of
##       the behavior of is(7, "integer"))
##       R-help (September 2008)
## 
## 
## Shan-Ho Chou: I got a question about eigenvector. I've tried input a symmetric
## matrix to both R (using eigen function) and Minitab, but the result is really
## different. Can anyone tell me what's wrong with that?
## Bill Venables: Simple. Minitab must be broken. Have you reported it to them?
##    -- Shan-Ho Chou and Bill Venables
##       R-help (July 2007)
## 
## 
## I still don't know what to do about the compromise between how statistics
## should be done and how journal editors seem to insist it should be done ...
##    -- Ben Bolker
##       R-sig-mixed-models (October 2008)
## 
## 
## Nothing exceeds like Excel.
##    -- Roger Koenker (in reply to a question why it takes so long to paste
##       19,000 lines of R code from Excel into R)
##       R-help (October 2008)
## 
## 
## Fabio Mulazzani: I need to obtain all the 9.somethingExp157 permutations that
## can be given from the numbers from 1 to 100.
## Ted Harding: To an adequate approximation there are 10^158 of them. Simply to
## obtain them all (at a rate of 10^10 per second, which is faster than the CPU
## frequency of most desktop computers) would take 10^148 seconds, or slightly
## longer than 3*(10^140) years. Current estimates of the age of the Universe are
## of the order of 1.5*(10^10) years, so the Universe will have to last about
## 2*(10^130) times as long as it has already existed, before the task could be
## finished. So: why do you want to do this?
##    -- Fabio Mulazzani and Ted Harding
##       R-help (November 2008)
## 
## 
## Cross posting is sociopathic.
##    -- Roger Koenker
##       R-help (November 2008)
## 
## 
## Some people familiar with R describe it as a supercharged version of
## Microsoft's Excel spreadsheet software.
##    -- Ashlee Vance (in his article "Data Analysts Captivated by R's Power")
##       The New York Times (January 2009)
## 
## 
## I think [R] addresses a niche market for high-end data analysts that want free,
## readily available code. [...] We have customers who build engines for aircraft.
## I am happy they are not using freeware when I get on a jet.
##    -- Anne H. Milley (director of technology product marketing at SAS, quoted
##       in Ashlee Vance's article "Data Analysts Captivated by R's Power")
##       The New York Times (January 2009)
## 
## 
## It's interesting that SAS Institute feels that non-peer-reviewed software with
## hidden implementations of analytic methods that cannot be reproduced by others
## should be trusted when building aircraft engines.
##    -- Frank Harrell (in response to the statement of the SAS director of
##       technology product marketing: "We have customers who build engines for
##       aircraft. I am happy they are not using freeware when I get on a jet.")
##       R-help (January 2009)
## 
## 
## This sounds like you have discovered homeopathic properties in Sweave! It will
## be serious if input files remember errors even after they have been removed.
##    -- Duncan Murdoch (after a user reported that two supposedly identical Rnw
##       documents work and lead to an error, respectively)
##       R-help (January 2009)
## 
## 
## Many problems come as a result of users forgetting that they are smarter than
## the computer. I see 3 ways to remedy the problem:
## 1. Make computers that are as smart or smarter than people.
## 2. Make the programmers anticipate every way that someone may use a particular
## function and make them implement all of the functionality even if they don't
## think it is worth the time/effort since there is an easy work around for many
## of the less likely used features.
## 3. Don't expect the computer to guess correctly and tell it exactly what you
## want it to do.
##    -- Greg Snow (in response to the question why as.Date() would not parse
##       "13/1/2001" in 'the only reasonable way')
##       R-help (January 2009)
## 
## 
## One of the beauties of S and R is the syntax closeness to mathematical
## notation. Many of us know that Lisp has beauties that S can never have, but
## that's really in different beauty-space.
##    -- Martin Maechler
##       R-devel (February 2009)
## 
## 
## R is Open Source and so you can modify it to emulate the bugs in other
## software: that is not one of the aims of its developers so please don't expect
## us to do so for you.
##    -- Brian D. Ripley (answering a request for a change to Excel-like non IEC
##       60559 standard conform rounding)
##       R-help (March 2009)
## 
## 
## The combination of some data and an aching desire for an answer does not ensure
## that a reasonable answer can be extracted from a given body of data.
##    -- John W. Tukey (the first of six "basics" against statistician's hubrises)
##       in "Sunset Salvo", The American Statistician 40(1), 72-76 (February 1986)
## 
## 
## I must say I feel very uneasy about giving advice on how to use R as a
## pre-processor for something that is to be done in Excel, but we are a broad
## church here!
##    -- Bill Venables
##       R-help (March 2009)
## 
## 
## No matter how much progress is made by the developers of screen readers somehow
## PDF remains less than accessible. It's a bit like a failed relationship in
## which one partner tries to win the other back with expensive gifts.
##    -- Paul R. Stanley (in a discussion about text-to-speech processors for the
##       benefit of blind users of LaTeX)
##       texhax mailing list (March 2009)
## 
## 
## As the name says, studentizing is left as an easy student exercise.
## Destudentizing is more difficult and usually called professorizing.
##    -- Dieter Menne (in response to a question whether residuals from nls() are
##       studentized)
##       R-help (April 2009)
## 
## 
## Hmm. I don't ever use R-squared, so I didn't implement it.
##    -- Thomas Lumley (in reply to a question about how to compute R-squared
##       using biglm)
##       R-help (April 2009)
## 
## 
## The issue really comes down to the fact that the questions: "exactly normal?",
## and "normal enough?" are 2 very different questions (with the difference
## becoming greater with increased sample size) and while the first is the easier
## to answer, the second is generally the more useful one.
##    -- Greg Snow (answering a question about a "normality test" suitable for
##       large data)
##       R-help (April 2009)
## 
## 
## Trellis graphics are a bit like hash functions: you can be close to the target,
## but get a far-off result.
##    -- Dieter Menne (about problems with creating a suitable lattice panel
##       function)
##       R-help (August 2008)
## 
## 
## The only people who should use the assign function are those who fully
## understand why you should never use the assign function.
##    -- Greg Snow
##       R-help (July 2009)
## 
## 
## 4 bits are not enough for me!
##    -- Philippe Grosjean (after conference dinner incl. wine in a conversation
##       about the representation of data in the ff package)
##       useR! 2009, Rennes (July 2009)
## 
## 
## Kelvin Lam: My institute has been heavily dependent on SAS for the past while,
## and SAS is starting to charge us a very deep amount for license renewal. Since
## we are a non-profit organization that is definitely not sustainable. The team
## is brainstorming possibility of switching to R, at least gradually. I am
## talking about the entire institute with considerable number of analysts using
## SAS their entire career. There's a handful of us using R regularly. What kind
## of problems and challenges have you faced?
## Frank Harrell: One of your challenges will be that with the increased
## productivity of the team you will have time for more intellectually challenging
## problems. That frustrates some people.
##    -- Kelvin Lam and Frank Harrell
##       R-help (July 2009)
## 
## 
## Yes we CRAN!
##    -- Aurelien Latouche (suggestion for a T-shirt slogan)
##       private communication (July 2009)
## 
## 
## Dear Uwe, thank you very much for your unvaluable time and effort.
##    -- Javier Cano (thanking Uwe Ligges for solving a coding problem)
##       R-help (July 2009)
## 
## 
## If I were to be treated by a cure created by stepwise regression, I would
## prefer voodoo.
##    -- Dieter Menne (in a thread about regressions with many variables)
##       R-help (October 2009)
## 
## 
## If one doesn't understand the output of lm(), then one's knowledge of
## statistics is insufficient to warrant using lm().
##    -- Peter Ehlers
##       R-help (October 2009)
## 
## 
## If 'fools rush in where angels fear to tread', then Bayesians 'jump' in where
## frequentists fear to 'step'.
##    -- Charles C. Berry (about Bayesian model selection as an alternative to
##       stepwise regression)
##       R-help (November 2009)
## 
## 
## You are going to need a few more sessions for this therapy to be useful, we
## cannot read your computer's mind, just interpret what you tell us. I don't want
## to jump ahead, but can you tell us the childhood of your crashes, etc.
##    -- Romain Francois (in a 'psychotherapy' session trying to understand the
##       sources of a useR's problems)
##       R-help (November 2009)
## 
## 
## It currently works (because I can't figure out how to make it an error) but you
## really should not do it.
##    -- Hadley Wickham (discussing a user's ggplot experiences)
##       R-help (October 2009)
## 
## 
## There's an informal tradition that those announcements [about R releases]
## contain at least one mistake, but apparently I forgot this time, so users have
## to make up their own....
##    -- Peter Dalgaard (about an apparent non-bug report in his former R-announce
##       message)
##       R-help (December 2009)
## 
## 
## It appears to me that you do not understand even basic statistics. As a
## corollary to that impression I would suggest that giving you advice about the
## use of R for scientific investigation could be morally similar to giving you
## advice about how to do your own household wiring.
##    -- David Winsemius
##       R-help (November 2009)
## 
## 
## Kakali Das: Is there any certification courses available in statistics in India
## which can be done after completing M.Sc. in Statistics? I need to know some
## courses which will be a specialization and can be done in distance mode from
## India?
## Kjetil Brinchmann Halvorsen: A specialization course (maybe better named
## generalization course) suitable to everybody, very cheap, and of immense
## quality, is to participate in the mailing list R-help, and simply try to
## understand (and then respond usefully) to questions arising there.
##    -- Kakali Das and Kjetil Brinchmann Halvorsen
##       Yahoo! Statisticians_group (November 2009)
## 
## 
## In general, it's much easier to create output from a R object than create an R
## object from output.
##    -- Hadley Wickham
##       R-help (December 2009)
## 
## 
## As Obi-Wan Kenobi may have said in Star Wars: "Use the source, Luke!"
##    -- Barry Rowlingson (answering a question on the documentation of some
##       implementation details)
##       R-devel (January 2010)
## 
## 
## Finally, a word of wisdom from a long-ago engineering colleague: "Whenever I
## see an outlier, I'm never sure whether to throw it away or patent it."
##    -- Berton Gunter (on outlier identification)
##       R-help (December 2009)
## 
## 
## As one of the developers of the nls function I would like to state that the
## lack of automatic ANOVA, R^2 and adj. R^2 from nls is a feature, not a bug :-)
##    -- Douglas Bates (in reply to a request for automatic ANOVA of NLS models as
##       in Statistica)
##       R-help (August 2000)
## 
## 
## My best advice regarding R^2 statistics with nonlinear models is, as Nancy
## Reagan suggested, "Just say no.".
##    -- Douglas Bates
##       R-help (August 2000)
## 
## 
## I recently attended a Ph.D. prelim exam where the candidate had proposed
## research on various ways of defining an R^2 statistic in the original data
## scale from a linear model fit to data in a transformed scale determined by the
## Box-Cox method. There were seven different possible definitions for R^2, all of
## which, as acknowledged by the candidate and by the thesis advisor, were
## incorrect. The purpose of this path-breaking study is to determine in some way
## which of these seven possible incorrect definitions should be used. My
## suggestion that the obvious answer was "none of them" was regarded as somewhat
## heretical.
##    -- Douglas Bates
##       R-help (August 2000)
## 
## 
## [Listing original copyright holders in R packages is] especially useful if
## there are areas of doubt and uncertainty (eg code published on ftp sites before
## people worried about licenses), since it at least gives you rigidly defined
## areas of doubt and uncertainty.[5]
## [5] Adams D, (1978) Hitchhiker's Guide. BBC Radio.
##    -- Thomas Lumley (in a discussion about copyright vs. licenses)
##       R-devel (January 2010)
## 
## 
## The widespread use of spreadsheets or SPSS data sets or SAS data sets which
## encourage the "single table with a gargantuan number of columns, most of which
## are missing data in most cases" approach to organization of longitudinal data
## is regrettable.
##    -- Douglas Bates (in a discussion about software for hierarchical data sets)
##       R-help (February 2010)
## 
## 
## R is a complex program. If you can't work out how to find the documentation, I
## doubt you will be able to use R.
##    -- David Kirby (in response to a vague question on where to find
##       documentation for R)
##       R-help (February 2010)
## 
## 
## If your ultimate interest is in real scientific progress, I'd suggest that you
## ignore that sentence (and any conclusion drawn subsequent to it).
##    -- Andy Liaw (in response to a question on the meaning of the sentence:
##       'Independent variables whose correlation with the response variable was
##       not significant at 5% level were removed')
##       R-help (March 2010)
## 
## 
## I begin to think that R needs a _mis_fortunes package...
##    -- Stephen L.R. Ellison
##       R-help (March 2010)
## 
## 
## I think what we are seeking is the marginal variance-covariance matrix of the
## parameter estimators (marginal with respect to the random effects random
## variable, B), which would have the form of the inverse of the crossproduct of a
## (q+p) by p matrix composed of the vertical concatenation of - L^{-1}RZX RX^{-1}
## and RX^{-1}. (Note: You do *not* want to calculate the first term by inverting
## L, use solve(L, RZX, system = "L") - [...] don't even think about using
## solve(L) - don't!, don't!, don't! - have I made myself clear? - don't do that
## (and we all know that someone will do exactly that for a very large L and then
## send out messages about "R is SOOOOO SLOOOOW!!!!" :-) )
##    -- Douglas Bates
##       R-SIG-Mixed-Models (March 2010)
## 
## 
## Only with a very high signal to noise ratio (e.g., high true R^2) can torturing
## data lead to a confession to something other than what the analyst wants to
## hear.
##    -- Frank Harrell
##       R-help (April 2010)
## 
## 
## If you don't want informative help files, it's really not much work to make
## uninformative ones.
##    -- Duncan Murdoch (in response to a question how to install a package
##       without editing .Rd files appropriately before)
##       R-help (May 2010)
## 
## 
## Dave Lubbers: R 2.7.2 - the manual says configure, make, which is what I did.
## So I did read the manual and followed the directions. The manual is too terse
## to get me there.
## Duncan Murdoch: You used the wrong tense. In referring to R 2.7.2, only past
## tenses are grammatically correct. If you want to say the manual *is* too terse,
## then you need to install R 2.11.0.
##    -- Dave Lubbers and Duncan Murdoch (on where to find libRmath)
##       R-devel (May 2010)
## 
## 
## Soham: How to compute the p-value of a statistic generally?
## Berton Gunter: runif(1)
##    -- Soham and Berton Gunter
##       R-help (May 2010)
## 
## 
## Karl Ove Hufthammer: Why not use 'predict.loess' (i.e., 'predict' on a loess
## object) directly?
## Greg Snow: Because that would be simple, straightforward, and make sense, and
## not require knowledge about less obvious functions.
##    -- Karl Ove Hufthammer and Greg Snow
##       R-help (May 2010)
## 
## 
## This reminds me of a quote I saw once (I think it may have been in one of those
## Murphy's Laws calendars), my parahpase:
## 
## If you make someone think that they are thinking,
##       They will love you for it.
## If you make them actually think,
##       They will hate you for it.
## 
## This explains why people love pie charts and hate more effective graphs.
##    -- Greg Snow
##       R-help (December 2009)
## 
## 
## On average, any data manipulation that can be described in a sentence or two of
## English can be programmed in one line in R. If you find yourself writing a long
## 'for' loop to do something that sounds simple, take a step back and research if
## an existing combination of functions can easily handle your request.
##    -- Erik Iverson
##       R-help (June 2010)
## 
## 
## There are actual error messages, and until you show them, we can not help as
## the mind reading machine is currently off for repairs.
##    -- Dirk Eddelbuettel (after reports about errors with R CMD check)
##       R-help (July 2010)
## 
## 
## Knut Krueger: Is there any function available to combine those p values?
## Stephan Kolassa: ?"+"
##    -- Knut Krueger and Stephan Kolassa (about ways to combine p values from
##       different tests)
##       R-help (July 2010)
## 
## 
## So, with the item in the original query:
## 
##   if (x<-3) do_something;
## 
## if they wrote it as
## 
##   if (x<(-3)) do_something;
## 
## there would be no problem (and no doubt about what went with what). Of course
## in complicated expressions this could induce an episode of ocular lispopia, but
## apart from that it's safe!
##    -- Ted Harding (about potential confusion between assignment and comparisons
##       with negative numbers)
##       R-devel (August 2010)
## 
## 
## Spencer Graves: What do you think about adding a "No RTFM" policy to the R
## mailing lists?
## Michael Dewey: You raise an interesting point but the responses to your post
## remind us that people (and indeed whole cultures) are not all situated at the
## same point on the continuum of directness between "It's a cow, stupid" and
## "From this side it looks not unlike a cow".
##    -- Spencer Graves and Michael Dewey (reply after a long discussion of a
##       potential "No RTFM" policy)
##       R-help (August 2010)
## 
## 
## Dirk Eddelbuettel: Could it get any more un-scientific and un-empirical? Maybe
## we should debate whether it is faster on Thursdays than on Wednesdays too?
## Thomas Lumley: That's easy. It's faster on Wednesdays, because faculty meeting
## and seminar are on Thursday (so more faculty are on campus using the computers)
## and homework for the applied regression sequence is due on Friday (so more
## students are using the computers).
##    -- Dirk Eddelbuettel and Thomas Lumley (after a rather vague disucssion
##       which language is fastest for numerical computations)
##       R-help and private communication (September 2010)
## 
## 
## datayoda: Bing is my friend...I found the cumsum() function.
## Dirk Eddelbuettel: If bing is your friend, then rseek.org is bound to be your
## uncle.
##    -- datayoda and Dirk Eddelbuettel (after searching for a function that
##       computes cumulative sums)
##       stackoverflow.com (October 2010)
## 
## 
## This is one of those cases, in my opinion, where R's documentation drops you
## into a flat landscape, in the middle of nowhere, in a thick mist.
##    -- Ted Harding (helping a user confused about R's partial matching behavior)
##       R-help (November 2010)
## 
## 
## I used a heuristic... pulled from my posterior. That makes it Bayesian, right?
##    -- JD Long (in a not too serious chat about modeling strategies)
##       Stackoverflow (November 2010)
## 
## 
## If you can do with[out a] Makefile, do without a Makefile.
##    -- Dirk Eddelbuettel (in a discussion about the benefits of (not) changing
##       the compiler in R packages)
##       R-SIG-HPC (December 2010)
## 
## 
## We also have huge amounts of work going into calibration transfer, i.e. making
## quantitative predictive models work on a different instrument. This is always a
## whole lot of work, and for some fields of problems at the moment considered
## basically impossible even between two instruments of the same model and
## manufacturer.
##    -- Claudia Beleites (in a discussion about publication bias and scientific
##       validity)
##       R-help (January 2011)
## 
## 
## This reminds me of a famous FORTRAN code snippet:
## 
## 10 STOP
## STOP
## STOP
## ! IN CASE STILL SKIDDING
## GOTO 10
##    -- Carl Witthoft (in response to the question how to completely stop R
##       scripts after stop())
##       R-help (January 2011)
## 
## 
## There is a reason that the speedometer in your car doesn't just read "slow" and
## "fast".
##    -- Frank Harrell (warning about the use of cutoffs after logistic regression)
##       R-help (February 2011)
## 
## 
## The style for JSS now enforces (via Achimitization) the use of quotes.
##    -- Michael Friendly (on whether to quote or not to quote arguments like
##       library("Hmisc") or help("mean"))
##       R-help (March 2011)
## 
## 
## Although there is almost never a "No" in R, the best short answer is: "No".
##    -- Uwe Ligges (in response to the question whether R has a concept of "const
##       objects" as in C++ that avoid modification)
##       R-help (March 2011)
## 
## 
## Tested solutions offered when reproducible examples are provided.
##    -- David Winsemius (suggesting a potential solution to a vague problem
##       description)
##       R-help (April 2011)
## 
## 
## The good way to do it is to include the following comment at the beginning:
## # This is a holy Script, please edit it not
##    -- Kenn Konstabel (on "... how to protect R Script files from inadvertent
##       editing by users.")
##       R-help (April 2011)
## 
## 
## Sorry, but this is the R-help list.  The 'R Spam' list is over at /dev/null.
##    -- Dirk Eddelbuettel (after a poster misused the mailing list for
##       advertising)
##       R-help (April 2011)
## 
## 
## You might also consider a more informative subject line. "Help needed" is true
## for all questions (not answers) on R-help. Just think, why is this list called
## R-***help***?
##    -- Peter Ehlers (after answering a help request)
##       R-help (April 2011)
## 
## 
## I think this is kind of like asking "will your Land Rover make it up my
## driveway?", but I'll assume the question was asked in all seriousness.
##    -- Ista Zahn (in response to a request for replication of some data
##       preprocessing done in SAS)
##       R-help (April 2011)
## 
## 
## In those first years after Bell Labs it was always a struggle to maintain my
## access to S and Port -- actually even Unix was a problem. I recall the director
## of campus computing services telling me that Unix wasn't appropriate for
## educational institutions since it was "too flexible."
##    -- Roger Koenker (in a discussion about nlminb)
##       private communication (April 2011)
## 
## 
## Joshua Wiley: ... the advantages of formal classes seem worth at least not
## entirely dismissing.
## Jim Lemon: Hmmm, yeah, that's about the grammatical equivalent of S4 classes.
##    -- Joshua Wiley and Jim Lemon (in a discussion about the relative advantages
##       of S3 and S4 classes)
##       R-help (May 2011)
## 
## 
## For recursive objects, search for recursive objects.
##    -- Barry Rowlingson (in a thread about "Recursive objects" and how to search
##       for former discussions about the topic)
##       R-devel (May 2011)
## 
## 
## Most people pay infinitely more to Microsoft for Windows than they pay to R
## Core for R. I hope that's also the ratio of their complaints to Microsoft about
## this bug to their complaints to us about R.
##    -- Duncan Murdoch (after a report that the default R version associated with
##       .R files could not be changed in the standard Windows dialogs)
##       R-help (May 2011)
## 
## 
## The trouble with nonstandard evaluation is that it doesn't follow standard
## evaluation rules...
##    -- Peter Dalgaard (about nonstandard evaluation in the curve() function)
##       R-help (June 2011)
## 
## 
## Vaishali Sadaphal: I need to give my R code to my client to use. I would like
## to protect the logic/algorithms that have been coded in R. This means that I
## would not like anyone to be able to read the code.
## Duncan Murdoch: R is an open source project, so providing ways for you to do
## this is not one of our goals. [...] If you think your client will steal from
## you, then you should find another client.
##    -- Vaishali Sadaphal and Duncan Murdoch
##       R-help (July 2011)
## 
## 
## Well..... SAS is SAS, but R is FREE...
##    -- Abhishek Rathore (in a discussion about R vs. SAS)
##       LinkedIn Group Stat-Math Statistics (July 2011)
## 
## 
## It is becoming apparent that you do not know how to use the results from either
## system. The progress of science would be safer if you get some advice from a
## person that knows what they are doing.
##    -- David Winsemius (in response to a user that obtained different linear
##       regression results in R and SPSS and wanted to know which one to use)
##       R-help (July 2011)
## 
## 
## For quite a while, bug-for-bug compatibility with S-PLUS v 3.x was considered
## important to allow people to port their packages between systems.
##    -- Peter Dalgaard
##       R-help (February 2009)
## 
## 
## I've been reading Data Analysis Using Regression and Multilevel/Hierarchical
## Models by Gelman and Hill. I'm getting acquainted with my new cartoon character
## friends Elmer (lmer) and Bugs, and at some point it may become clear what is
## going on. At the moment though, the whole thing seems a bit Looney Tunes.
##    -- Paul Miller
##       R-SIG-Mixed-Models (August 2011)
## 
## 
## Thank your for your entry in the Poorly Capitalized and Inadequately Searched
## Posting Contest. You will be advised of your ranking in due course.
##    -- David Winsemius (in response to an all-lower-case question on a simple
##       clustering task)
##       R-help (August 2011)
## 
## 
## Don't do as I say, do as Hadley does.
##    -- Barry Rowlingson (in a discussion about the workflow for writing R
##       packages, see also fortune(128))
##       R-devel (September 2011)
## 
## 
## R is wonderful, but it cannot work magic.
##    -- Rolf Turner (answering a request for automatic generation of 'data from a
##       known mean and 95% CI')
##       R-help (October 2011)
## 
## 
## If we told you all of R's secrets, we'd have to kill you.
##    -- Duncan Murdoch (answering a request for too detailed documentation)
##       R-help (October 2011)
## 
## 
## I know barely more than zero about R: until yesterday I didn't know how to
## spell it.
##    -- Pete Wilson
##       stackoverflow (October 2011)
## 
## 
## Kevin Wright: Doug already mentioned the limitations of open source
## optimization code. Open source has benefits, but simply being open doesn't
## necessarily mean better.
## 
## Joshua Wiley: Actually, a series of studies I have conducted including
## theoretical proofs and empirical evidence proves open source to be
## unequivocally better. If you are interested, you can pay me to use the
## assertion, but the supporting details are restricted so you must accept that my
## internal checks and quality control are sufficient and the findings are correct
## and accurate. Details follow.
## 
## License and Use of I. Ben Fooled's findings:
## 
## "Open source is unequivocally better" is $10 for personal use or $20 for each
## public use of the assertion. It is also possible to get a yearly license, but
## note that this does not include updates. Support is extra. If you wish to use
## the assertions on a multicore system, there is an additional $5 per core per
## assertion fee. A student version is available for pedagogical purposes only:
## "pn src s nqvcll bttr". If you need more than 20 letters, please upgrade to the
## full version.
## 
## This is the best thing since bread, but don't take my word for it. Here are
## some trite testimonials putatively from customers:
## 
## "Using I. Ben Fooled's assertion has really revolutionized my work. I love it.
## I use it every day" ~ Jane, Data Expert
## 
## "I lost 130 lbs on this system", 'wait, this isn't a weightloss testimonial?
## Sorry!' "My work is easily twice as efficient since I switched to I. Ben
## Fooled's product." ~ John, Entrepeneur
##    -- Kevin Wright and Joshua Wiley
##       R-SIG-Mixed-Models (October 2011)
## 
## 
## As with many tools in this domain, for effective use they require more
## knowledge than many of their users possess, and can be dangerous because they
## seem to "work".
##    -- John C. Nash (in the context of stochastic optimization)
##       R-help (December 2011)
## 
## 
## Unfortunately, this is one of those situations where as far as I can tell all
## of the real statisticians are out there playing with large data sets where the
## small-sample corrections are not so important and leaving the rest of us to
## figure it out for ourselves...
##    -- Ben Bolker (about the 'correct' degrees-of-freedom correction in Wald
##       statistics for GLMMs)
##       R-SIG-Mixed-Models (December 2011)
## 
## 
## David Winsemius: Any of your staff can view the (open) source code just as any
## other member of the human race.
## Uwe Ligges: I do not think the GPL excludes other species...
##    -- David Winsemius and Uwe Ligges (following a request for access to the R
##       sources)
##       R-help (December 2011)
## 
## 
## I think the amount of people on this list who understand your question is
## roughly zero.
##    -- Uwe Ligges (in response to a non-reproducible problem for which
##       additionally subject and body did not match well)
##       R-help (January 2012)
## 
## 
## It's not meant for sampling weights. It's meant for precision weights. How best
## to include sampling weights in mixed models is a research problem at the
## moment, but you can rely on getting the wrong answer if you just use the
## weights= argument.
##    -- Thomas Lumley (in reponse to a question about specifying sampling weights
##       in lme4)
##       R-help (January 2012)
## 
## 
## Perhaps same thing that will happen when those monks finish the Towers of Hanoi?
## 2 * 3^198
## [1] 5.902533e+94
##    -- David Winsemius (in reponse to a question about expand.grid(x1=1, x2=1:2,
##       x3=1:3, x4=1:3, x5=1:3, ..., x200=1:3))
##       R-help (January 2012)
## 
## 
## I'll go back to lurking in the daily R-Helps and not ask any more questions
## until I've read all the old R-help messages. I'm working on December 1998 right
## now and reading forward. Perhaps by next year I'll will have read all the old
## R-help postings and I'll dare ask another question then.
##    -- Earl F. Glynn
##       R-help (January 2012)
## 
## 
## Dirk Eddelbuettel: Devel is in the detail.
## Douglas Bates: Typo or profound insight?
##    -- Dirk Eddelbuettel and Douglas Bates
##       Rcpp-devel (February 2012)
## 
## 
## There seem to be three canons: R-Forge, Rforge, and for the Followers of
## Wickham, github. Four, if you count OmegaHat.
##    -- David Winsemius (in a discussion about the 'canonical' source code
##       management platform for R)
##       R-help (February 2012)
## 
## 
## The problem here is that the $ notation is a magical shortcut and like any
## other magic if used incorrectly is likely to do the programmatic equivalent of
## turning yourself into a toad.
##    -- Greg Snow (in response to a user that wanted to access a column whose
##       name is stored in y via x$y rather than x[[y]])
##       R-help (February 2012)
## 
## 
## Etienne B. Racine: I haven't seen any documentation about this behavior.
## Edzer Pebesma: Isn't the more important question whether you saw documentation
## that is in conflict with this behaviour?
## Tom Gottfried: Taking that to the extremes: undocumented software always works
## perfectly.
##    -- Etienne B. Racine, Edzer Pebesma, and Tom Gottfried
##       R-SIG-Geo (February 2012)
## 
## 
## That's a casual model, not a causal model - you can tell the difference by
## looking for the word "excel".
##    -- Hadley Wickham (commenting on an Excel chart showing student's SAT score
##       increases with family income, without considering further covariates)
##       http://twitter.com/#!/hadleywickham (February 2012)
## 
## 
## Careful now, or I'll get into a rant about how much you pay for ArcGIS and
## whether you are getting value for money if bugs become 'well-known'! Bugs in R
## tend to be either 'obscure' or 'fixed' :)
##    -- Barry Rowlingson (in response to a request for rgdal to work around a
##       well-known bug in ArcGIS)
##       R-SIG-Geo (March 2012)
## 
## 
## Igor Sosa Mayor: The problem is that it gives the result that I want.
## Sarah Goslee: That's a new sort of problem.
##    -- Igor Sosa Mayor and Sarah Goslee
##       R-help (March 2012)
## 
## 
## I'd have called them .Rob for the current .rds file format (single object), and
## .Ros for the current .rda file format (>1 objects). [actually I would have
## called them .Rob or .Robject, and .Rwk or .Rworkspace but then I noticed that
## if you had a .Rob you had to have a .Ros so as not to favour one or other of
## the project originators :)]
##    -- Barry Rowlingson (in a discussion about .rda and .rds formats)
##       R-help (April 2012)
## 
## 
## R is a strange, deeply flawed language that nevertheless has an enthusiastic
## and rapidly growing user base. What about R accounts for its popularity in its
## niche? What can language designers learn from R's success?
##    -- John Cook (Abstract of "Why and How People Use R")
##       Lang.NEXT 2012 (April 2012)
## 
## 
## Finally, as author of the code used to put CG in optim, I'll advise against its
## use. One of my least successful pieces of code. Rcgmin is better, but you
## really do need analytic derivatives to make it sing. For 5 parameters, use NM,
## or better the nmk from dfoptim package.
##    -- John C. Nash (in a discussion about optimization of a multinomial
##       log-likelihood)
##       R-help (May 2012)
## 
## 
## Some OSX users know that OSX is really Unix [...], others think that OSX is
## cooler Windows, and they have ontological problems with non-Apple phenomena and
## constructs.
##    -- Roger Bivand
##       R-SIG-Geo (May 2012)
## 
## 
## I see three drivers of problems with R.
## The first is that R is useful. [...]
## The second driver of problems with R is that it is both a programming language
## and an interactive language. There is a tension there that is unavoidable. [...]
## The third big driver of problems is that R is not software, it is a community.
##    -- Patrick Burns (Inferno-ish R)
##       CambR User Group Meeting, Cambridge (May 2012)
## 
## 
## As long as R stayed in Ross and Robert's computer lab, R could be perfectly
## clean and beautiful. [...]
## The moral of the story is that if you want to create a beautiful language, for
## god's sake don't make it useful.
##    -- Patrick Burns (Inferno-ish R)
##       CambR User Group Meeting, Cambridge (May 2012)
## 
## 
## If you think you can learn all of R, you are wrong. For the foreseeable future
## you will not even be able to keep up with the new additions.
##    -- Patrick Burns (Inferno-ish R)
##       CambR User Group Meeting, Cambridge (May 2012)
## 
## 
## The phrase "does not work" is not very helpful, it can mean quite a few things
## including:
## * Your computer exploded.
## * No explosion, but smoke is pouring out the back and microsoft's "NoSmoke"
## utility is not compatible with your power supply.
## * The computer stopped working.
## * The computer sits around on the couch all day eating chips and watching talk
## shows.
## * The computer has started picketing your house shouting catchy slogans and
## demanding better working conditions and an increase in memory.
## * Everything went dark and you cannot check the cables on the back of the
## computer because the lights are off due to the power outage.
## * R crashed, but the other programs are still working.
## * R gave an error message and stopped processing your code after running for a
## while.
## * R gave an error message without running any of your code (and is waiting for
## your next command).
## * R is still running your code and the time has exceeded your patience so you
## think it has hung.
## * R completed and returned a result, but also gave warnings.
## * R completed your command, but gave an incorrect answer.
## * R completed your command but the answer is different from what you expect
## (but is correct according to the documentation).
## 
## There are probably others. Running your code I think the answer is the last one.
##    -- Greg Snow
##       R-help (April 2012)
## 
## 
## Richard M. Heiberger: That will make it easier for the next person converting
## from windows to mac.
## Simon Urbanek: Well, it certainly doesn't make sense to introduce Windows
## idiosyncrasies to Mac - if you convert to Mac you will need to buy into the
## "just-works" philosophy. If you are trying to recreate your Windows
## environment, you have already failed and should go back :).
##    -- Richard M. Heiberger and Simon Urbanek (in a discussion about the
##       location of the R FAQ on different platforms)
##       R-SIG-Mac (July 2012)
## 
## 
## I think, therefore I R.
##    -- William B. King (in his R tutorials)
##       http://ww2.coastal.edu/kingw/statistics/R-tutorials/ (July 2010)
## 
## 
## Jeremy Koster: My students were looking at the estimated varying intercepts for
## each higher-level group (or the "BLUP's", as some people seem to call them).
## Douglas Bates: As Alan James once said, "these values are just like the BLUPs -
## Best Linear Unbiased Predictors - except that they aren't linear and they
## aren't unbiased and there is no clear sense in which they are "best", but other
## than that ..."
##    -- Jeremy Koster and Douglas Bates
##       R-SIG-Mixed-Models (July 2012)
## 
## 
## I wish there was a better way for me to learn this stuff than constantly being
## wrong in public.
##    -- Hadley Wickham (in a discussion about efficient expansion of vectors and
##       algorithmic complexity)
##       R-help (July 2012)
## 
## 
## Hans Ole Orka: I try to reproduce the SAS proc reg stepwise model selection
## procedure in R.
## Brian D. Ripley: But why? If you want 1950s statistical methods why not use a
## 1960s package? There are enough problems with stepwise selection (see e.g. the
## book by Frank Harrell and many postings here) even with a well-defined
## criterion like AIC, but that is better than an ad hoc algorithm, especially one
## based on forwards selection.
##    -- Hans Ole Orka and Brian D. Ripley
##       R-help (September 2007)
## 
## 
## The computational ease with which an abundance of parameters can be estimated
## should not be allowed to obscure the probable unwisdom of such estimation from
## limited data.
##    -- Arthur P. Dempster
##       in "Covariance selection", Biometrics 28 (1), 157-175 (March 1972)
## 
## 
## I thought RStudio was amazing, but RStudio with knitr is approximately
## `formatC(round(runif(1, 1e8, 1e9)), digits=10, big.mark=',')` times better than
## RStudio alone!
##    -- @Geneorama (comment on RStudio's new web publishing service)
##       RStudio blog (June 2012)
## 
## 
## Awkwardly worded questions will get much better answers if they are accompanied
## by some test data.
##    -- David Winsemius (in response to a poster's apology for an 'awkwardly
##       worded question')
##       R-help (September 2012)
## 
## 
## Liang Che: For example, if coefficient's p-value is less than 0.1 I want the
## stepwise to automatically drop that variable. Can the stepAIC be customized to
## do that? SAS seems to be able to customized stepwise function with p-value or
## cooks'd.
## David Winsemius: You might take some time to ponder the possibility that the
## fact that it's not easy in R might be useful information in its own right.
##    -- Liang Che and David Winsemius
##       R-help (September 2012)
## 
## 
## Josh O'Brien: Figuring out what was going on here obsessed me for several days,
## before I realized "Oh, duh. I should ask this on [Stackoverflow] and get some
## better minds working on this for me!"
## Brandon Bertelsen: I've learned to give up after 30 minutes. It's cheaper to
## pay people in checkmarks and upvotes :)
##    -- Josh O'Brien and Brandon Bertelsen (discussing a package's definition of
##       reorder.factor changing results)
##       Stackoverflow (October 2012)
## 
## 
## This should be FAQ 0.0. No other thing is asked as frequently as this. This is
## the FAQest of all FAQs, and a mother of all FAQs.
##    -- Jari Oksanen (answering yet another question related to FAQ 7.31)
##       R-help (February 2013)
## 
## 
## I've already got an apartment reserved for me in one of Pat Burns's "R Inferno"
## levels, and I don't want to descend even further.
##    -- Berton Gunter (commenting on his own somewhat off-topic comment)
##       R-help (February 2013)
## 
## 
## Running as administrator is like heroin... any problems it solves it replaces
## with worse problems.
##    -- Jeff Newmiller (about running R CMD check as administrator)
##       R-help (February 2013)
## 
## 
## I think the rule is that you can do anything as long as you don't complain. If
## you want to complain, you must follow the instructions.
##    -- Jari Oksanen (about not following the instructions when compiling R-devel)
##       R-devel (February 2013)
## 
## 
## I think that the formula language does allow expressions with '/' to represent
## nested factors but I can't check right now as there is a fire in the building
## where my office is located. I prefer to simply write nested factors as factor1
## + factor1:factor2.
##    -- Douglas Bates
##       R-SIG-Mixed-Models (February 2013)
## 
## 
## Bandwidth selection is an unresolved (and possibly unsolvable) problem in
## smoothing, so you're perfectly justified in trying/choosing an arbitrary value
## if it produces good pictures!
##    -- Adrian Baddeley (answering a user's question about the choice of
##       smoothing parameter when using the density.ppp() function from the
##       spatstat package)
##       private communication (March 2013)
## 
## 
## Rickyfox: Dang it how could I have missed this? Thx for the answer, I feel
## incredibly stupid now.
## Juba: Never underestimate the power of R to make you feel stupid.
##    -- Rickyfox and Juba (in an exchange over how ifelse() works)
##       stackoverflow.com (March 2013)
## 
## 
## If things are not readily available in R it is always good to pause and reflect
## if there might be a good reason.
##    -- Brian D. Ripley (about how to get t statistics for arima())
##       R-help (March 2013)
## 
## 
## Sooner or later most R beginners are bitten by this all too convenient
## shortcut. As an R newbie, think of R as your bank account: overuse of
## $-extraction can lead to undesirable consequences. It's best to acquire the
## '[[' and '[' habit early.
##    -- Peter Ehlers (about the use of $-extraction)
##       R-help (March 2013)
## 
## 
## Most optimization problems have more than one answer, and the "wrong" ones
## often seem to be easier to find.
##    -- John C. Nash (about nonlinear optimization problems)
##       R-help (April 2013)
## 
## 
## JPM Miao: Why can't R understand if(num!=NA)?
## Peter Dalgaard: Because comparison with an unknown value yields an unknown
## result.
## David Winsemius: Anything else would violate the Second Law of Thermodynamics.
## We cannot have comparisons reducing entropy, now can we? Uncertainty cannot run
## uphill.
##    -- JPM Miao, Peter Dalgaard, and David Winsemius (on why is.na() is needed)
##       R-help (May 2013)
## 
## 
## In that flash of darkness he heard people saying - he was a failure - that R
## was beyond him.
##    -- Virginia Woolf
##       To The Lighthouse (May 1927)
## 
## 
## This is a bit like asking how should I tweak my sailboat so I can explore the
## ocean floor.
##    -- Roger Koenker (in response to a question about tweaking the quantreg
##       package to handle probit and heckit models)
##       R-help (May 2013)
## 
## 
## Justin: Is there a function that just does whatever I'm thinking (aka whatever
## my homework question is...)?
## Joshua Ulrich: That's the magic_pony function.
##    -- Justin and Joshua Ulrich
##       stackoverflow.com (June 2013)
## 
## 
## The job of computer output, in my view, is to be as informative as possible
## while keeping the output as terse as possible.
##    --  John Maindonald
##       R-SIG-Mixed-Models (June 2013)
## 
## 
## You need to get the hang of reading the online help. The information required
## is actually there in ?dotchart --- it's just tersely and obscurely expressed. A
## certain degree of optimism is required. You need to ***believe*** that the
## information is there; then ask yourself "What could they possibly mean by what
## they have written that would tell me what I need to know?".
##    -- Rolf Turner (on reading the help pages)
##       R-help (June 2013)
## 
## 
## But a spatial benchmark suite would be nice anyway.
## I've just benchmarked my time and I have 0s to do this.
##    -- Barry Rowlingson (on benchmarking spatial functionality and assessing
##       improvements with pqR)
##       R-SIG-Geo (June 2013)
## 
## 
## John Fox: I've never understood why it's legal to change the built-in global
## "constants" in R, including T and F. That just seems to me to set a trap for
## users. Why not treat these as reserved symbols, like TRUE, Inf, etc.?
## Rolf Turner: I rather enjoy being able to set pi <- 3.
##    -- John Fox and Rolf Turner
##       R-help (June 2013)
## 
## 
## I have it on good authority that a few of the statements in Bolker (2008) are
## out of date or slightly incorrect :-) .
##    -- Ben Bolker (in reply to a user that cited his 2008 practical guide to
##       generalized linear mixed models)
##       R-SIG-Mixed-Models (August 2013)
## 
## 
## Well, the biggest room in the world is the room for improvement :)
##    -- Soren Hojsgaard ( in reply to a suggestion to make pbkrtest and lme4 more
##       robust)
##       R-SIG-Mixed-Models (August 2013)
## 
## 
## I'm still confused about how to avoid the wrath of the CRAN-devel daemon, whose
## appetite for new morsels of developer flesh seems ever increasing and makes
## keeping even a stable package up-to-date a moving target.
##    -- Michael Friendly (about checking packages for CRAN)
##       R-devel (September 2013)
## 
## 
## Maybe I'll become a theoretician. Nobody expects you to maintain a theorem.
##    -- Doug Bates (about keeping both Matrix and RcppEigen in sync with CHOLMOD)
##       lme4-authors (September 2013)
## 
## 
## Bert Gunter: However, I do not understand the substitute(...()) idiom. Would
## you care to explain it? (No is an acceptable answer!).
## Bill Dunlap: I don't completely understand it either, I treat it as an idiom. I
## saw it on this list once.
##    -- Bert Gunter and Bill Dunlap (on how to recover object names when using
##       the ... argument in a function)
##       R-help (October 2013)
## 
## 
## The existence of a method is not a sufficient reason to use that method.
##    -- Jari Oksanen (about relative advantages of several multivariate analysis
##       methods)
##       R-SIG-Ecology (November 2013)
## 
## 
## I would heed the warnings and diagnostics. They are there for a reason. The
## Ostrich algorithm does not help you.
##    -- Dirk Eddelbuettel (in reply to a questioner who said he had not run R CMD
##       check because he suspected other problems would be found)
##       Stackoverflow (November 2013)
## 
## 
## Either I am misunderstanding your intent or you need another cup of coffee.
##    -- Rolf Turner (in response to a user who did not understand his advice)
##       R-help (November 2013)
## 
## 
## You seem to be falling prey to a common misconception that "R" is some
## monolithic tool, when in fact it is a herd of cats.
##    -- Jeff Newmiller (in response to a user who did not look for packages)
##       R-help (December 2013)
## 
## 
## Basically, you are calling .Internal from the command line. It is not designed
## to be called from there and only wizards know what happens if it is. (The set
## of wizards who might know whether it makes any sense at all does not include
## me!)
##    -- Peter Dalgaard
##       R-devel (May 2014)
## 
## 
## Jeff Newmiller's rhelp/ggplot strategy checklist:
## > Reproducibility == good (except for mixing in invalid code and failing to set
## RNG seed)
## > HTML email == bad
## > Missing subject == bad
## > Using matrices with ggplot == pushing a rope
## > Using for loops to build a ggplot == missing the point
## > Using wide-form data with ggplot == pushing a rope
##    -- Jeff Newmiller (in response to a somewhat involved ggplot2 question)
##       R-help (June 2014)
## 
## 
## Michael Mason: Thanks! That worked.
## Martin Maechler: Of course: As in about 99.99% of all cases where Bill Dunlap
## helps.
##    -- Michael Mason and Martin Maechler (after Bill Dunlap helped with a
##       plot.hclust problem)
##       R-help (November 2014)
## 
## 
## When in doubt, keep adding slashes until it works.
##    -- Joran Elias (on how to escape a backslash in R)
##       Stackoverflow (March 2015)
## 
## 
## .RData files (the ones with nothing before the period) are just traps for your
## future self, with no documentation. I avoid them like the plague.
##    -- Jeff Newmiller (in a discussion on whether to store certain defaults in
##       .Rprofile or .RData files)
##       R-help (March 2015)
## 
## 
## hist(rnorm(1e6), col = colors()[grep("grey", colors())], nclass = 108, main =
## "R's 108 Shades of Grey")
##    -- Fritz Scholz (proving R has more than 50 Shades of Grey)
##       private communication (March 2015)
## 
## 
## In real life, be very careful before denigrating a child to its mother or
## father... so calling a something a bug in R which is none, is evoking feelings
## among R's parents... ;-)
##    -- Martin Maechler (after a package author reported a bug in R that was
##       actually a bug in his own code)
##       R-devel (May 2015)
## 
## 
## Welcome to the wonderful world of heterogeneity (tau^2) estimators in
## meta-analysis. The question of which estimator to choose never fails to provide
## fun and entertainment for the entire family and I predict that it will do so
## for generations to come. Please quote me on that!
##    -- Wolfgang Viechtbauer (in response to a request for the appropriate
##       heterogeneity estimator)
##       R-SIG-Mixed-Models (May 2015)
## 
## 
## Rahul: Microsoft's being a part of this makes me worried about R's future. It's
## like a Michelin star restaurant becoming part of a McDonald's franchise chain.
## Andrew Gelman: I dunno, R's not quite a Michelin-starred restaurant. It's more
## like a food truck.
## Daniel Lakeland: I would say if R is like a food truck, it's a TARDIS food
## truck that's bigger on the inside and offers a vast array of multi-ethnic
## cuisine which, if not all entirely tasty, is at least A rated on sanitation.
##    -- Rahul, Andrew Gelman, and Daniel Lakeland (in a discussion about Excel
##       interfaces and Microsoft being part of the new R Consortium)
##       AndrewGelman.com (July 2015)
## 
## 
## The urge to imitate other statistical packages that rely on profusion of
## dummies should be resisted. R repression functions can handle factor variables.
##    -- David (Muphry) Winsemius (about factors in regressions...and the dangers
##       of drinking and deriding)
##       R-help (September 2015)
## 
## 
## It will be included in the next version of devtools - it's totally do-able, but
## no one has done it yet.
##    -- Hadley Wickham (about doing things)
##       R-help (February 2016)
## 
## 
## Barry Rowlingson: I've been unable to locate any version of S online. Does
## anyone have a copy, somewhere, rusting away on an old hard disk or slowly
## flaking off a tape? [...] Obviously this would be for archaeological purposes.
## John Chambers: [...] somehow "historical" would be less unnerving than
## "archeological".
## Barry Rowlingson: At least I didn't say palaeontological.
##    -- Barry Rowlingson and John Chambers (on the availability of the source
##       code for early versions of S)
##       R-devel (February 2016)
## 
## 
## Ooops, that was answering the question you actually asked. The one you meant to
## ask is answered by this part:
## The sort order for character vectors will depend on the collating sequence of
## the locale in use: see Comparison.
##    -- Peter Dalgaard (demonstrating his mind-reading skills when answering
##       questions about sorting characters)
##       R-help (March 2016)
## 
## 
## Note that there is a reasonable discussion in Writing R Extensions, but as
## always, you have to find it first.
##    -- Dirk Eddelbuettel (in a discussion about where to include data-generating
##       scripts in an R package)
##       R-pkg-devel (March 2016)
## 
## 
## They're just statistics. How could it hurt to look at them?
##    -- Duncan Murdoch (in a discussion about the merits and perils of p-values)
##       R-help (April 2016)
## 
## 
## I prefer _Melodramatic Analysis_.
##    -- Dirk Eddelbuettel (after Barry Rowlingson noticed 'Sentimental Analysis'
##       as a misspelled 'Sentiment Analysis' on datascience.stackexchange.com)
##       R chatroom on stackoverflow.com (April 2016)
## 
## 
## Weird. Over at ESS they discuss how to debug with %>% when I always thought
## that having %>% was the bug.
##    -- Dirk Eddelbuettel (about an ESS issue ticket discussion concerning
##       debugging when %>% is used)
##       R chatroom on stackoverflow.com (April 2016)
## 
## 
## The problem with attach is what it does, not how it does it. If you do the same
## thing attach does any other way, you've managed to create all the problems of
## attach without using attach. If I tell you chopping your finger off with a
## knife is bad, your question shouldn't be "Okay, knives are bad. How do I chop
## my finger off without a knife? Can I use scissors or is there a better way?"
##    -- Gregor (about alternatives to attach())
##       stackoverflow.com (May 2016)
## 
## 
## Teach a man to fish, and he'll use StackOverflow for a day. Give him a fish,
## and he'll use StackOverflow for a lifetime of free fish.
##    -- Joshua Ulrich (about pointing to R documentation on StackOverflow)
##       stackoverflow.com (May 2016)
## 
## 
## You should try things; R won't break.
##    -- Duncan Murdoch (encouraging a user to explore R's graphical capabilities)
##       R-help (May 2016)
## 
## 
## Please don't apply R like a magic answers box, because you can mislead others
## and cause harm.
##    -- Jeff Newmiller (about how much statistical knowledge is needed for using
##       R)
##       R-help (May 2016)
## 
## 
## Your example requires all choices of 1541 out of 3000, which I would expect to
## take somewhere near age-of-the-universe seconds to compute. The code uses a
## clever nested computation due to Gail et al which will cut that time down to
## infinity/10.
##    -- Terry Therneau (on an example with exact computations in coxph)
##       R-help (June 2016)
## 
## 
## Performance tweaks are more often bug sources than fixes.
##    -- Duncan Murdoch (on Martin Maechler's optimistic plan of porting
##       performance improvements from R's development version to R x.y.z (z >= 1)
##       patch releases)
##       R-core (June 2016)
## 
## 
## This is like asking, "My car doesn't work. Can anyone tell me what is wrong?"
##    -- Jeff Newmiller (in response to the generic question why "some packages do
##       not work" in R 3.2.5)
##       R-help (June 2016)
## 
## 
## If we put in a function into rstan that dropped chains, people would use it.
##    -- Ben Goodrich (about (not) discarding selected chains from a stanfit
##       object)
##       Stan-users (December 2016)