Berton Gunter has written in part:> A few comments:> First, your remarks are interesting and, I would say, mainly wellfounded. However, I think they > are in many respects irrelevant, although they do point to the much bigger underlying issue,> which Roger Peng also hinted at in his reply.> I think they are sensible because R IS difficult; the documentation isoften challenging, which is> not surprising given (a) the inherent complexity of R; (b) thedifficulty in writing good> documentation, especially when many of the functions being documentedare inherently> technical, so subject matter knowledge (CS, statistics, numericalanalysis ,...) must be> assumed;My experience has been that the real challenge is not understanding the documentation, but finding it. Once I know the names of one or more candidate functions I am happily on my way. One of the delights of reading r-help is that one keeps discovering useful functions. In the best of all possible worlds I could ask an intelligent agent to summon up the k-nearest neighbor functions that would "do X." Not likely. Years ago StatSci Europe published a handy little "Complete Listing of S-PLUS Functions", categorized in some way. I found it useful. Something similar for R would not go amiss. I know, it would want to be 420 pages rather than 42. ********************************************************** Cliff Lunneborg, Professor Emeritus, Statistics & Psychology, University of Washington, Seattle cliff at ms.washington.edu
On Wed, 18 Aug 2004 12:16:21 -0700, "Cliff Lunneborg" <cliff at ms.washington.edu> wrote :> Years >ago StatSci Europe published a handy little "Complete Listing of S-PLUS >Functions", categorized in some way. I found it useful. Something >similar for R would not go amiss. I know, it would want to be 420 pages >rather than 42.The R Reference manual does this for the base packages. The HTML help pages come sort of close for other packages, though they're on a separate page for each package. If you really want it all in one place, it would presumably be fairly easy to modify the code that produces those and have everything appear on one big page, or write a script that glued together all the <RHOME>/library/html/00Index.html files. Duncan Murdoch
On Wed, 18 Aug 2004, Cliff Lunneborg wrote:> Berton Gunter has written in part: > > > A few comments: > > > First, your remarks are interesting and, I would say, mainly well > founded. However, I think they > are in many respects irrelevant, > although they do point to the much bigger underlying issue, > > which Roger Peng also hinted at in his reply. > > > I think they are sensible because R IS difficult; the documentation is > often challenging, which is > > not surprising given (a) the inherent complexity of R; (b) the > difficulty in writing good > > documentation, especially when many of the functions being documented > are inherently > > technical, so subject matter knowledge (CS, statistics, numerical > analysis ,...) must be > > assumed; > > My experience has been that the real challenge is not understanding the > documentation, but finding it. Once I know the names of one or more > candidate functions I am happily on my way. One of the delights of > reading r-help is that one keeps discovering useful functions. In the > best of all possible worlds I could ask an intelligent agent to summon > up the k-nearest neighbor functions that would "do X." Not likely.help.search does a better job than it is given credit for.> Years ago StatSci Europe published a handy little "Complete Listing of > S-PLUS Functions", categorized in some way. I found it useful. Something > similar for R would not go amiss. I know, it would want to be 420 pages > rather than 42.What is R in this context? There are several hundred addons on CRAN, BioC and elsewhere. R's HTML search or help.search will give you a complete listing over installed packages by `keyword', which is what the `Complete Listing of S-PLUS Functions' I saw was about. Windows users should try the full-text searches in CHM help, especially for package stats. The problem is to know what to search for. To pick a recent example, to use `logistic-normal model' for a random-intercept GLMM is not going to work, but Googling will usually bring up synonyms. -- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
On 08/18/04 12:16, Cliff Lunneborg wrote:>My experience has been that the real challenge is not understanding the >documentation, but finding it. Once I know the names of one or more >candidate functions I am happily on my way. One of the delights of >reading r-help is that one keeps discovering useful functions. In the >best of all possible worlds I could ask an intelligent agent to summon >up the k-nearest neighbor functions that would "do X."I have found the HtDig search engine at my site (accessible through "Search" on the left side of the main R page, or directly as http://finzi.psych.upenn.edu) to be pretty useful in this regard, although it is a long way from artificial intelligence, which would recognize similar meanings. It fails for me mostly when different disciplines have different names for the same thing. (Economists hate to admit that many of the statistical ideas they use were invented/discovered by psychologists.) That said, I'm thinking of switching to the Xapian search engine (http://www.redhat.com/archives/fedora-devel-list/2004-July/msg01576.html), and I would welcome any opinions about it. HtDig is a pain; only one version of it (an old one) seems to work on Fedora Core 2, and it now takes almost 10 hours to update each month on a very fast computer (Pentium 4 2.80GHz with Serial ATA disk controller). Jon -- Jonathan Baron, Professor of Psychology, University of Pennsylvania Home page: http://www.sas.upenn.edu/~baron