[R] Packages - a great resource, but hard to find the right one

[John Sorkin]

This thread has generated a number of comments. My reading of the
messages is
(1) There are a number of ways to find packages, no one is perfect.
(2) Packages vary in quality and a review process that would help
identify better packages and
suggest improvements for packages would be helpful. 

As far as finding packages goes, I believe that R would be a better
tool if a package finder were
built into the R distribution and added to the menus presented by the R
GUI. The package finder
would allow the user to enter search words and the package finder would
return a series of packages
along with brief précis to each package and a link to package reviews.
I know that this suggestion would
require a lot of work both by the R maintainers (to develop the search
system and make it a part of
the GUI), package developers (to add some form of metadata to there
packages) and the 
R community (to write the reviews), but I think it is important if R is
to continue to grow and develop.
Currently we have a wonderful platform which has grown tremendously
over the years, which is 
in danger of becoming less useful that it deserves to be because of a
lack of a centralized,
modern, easily searchable package finding system.

Thanks to all who contributed to the discussion.

John


John Sorkin M.D., Ph.D.
Chief, Biostatistics and Informatics
Baltimore VA Medical Center GRECC,
University of Maryland School of Medicine Claude D. Pepper OAIC,
University of Maryland Clinical Nutrition Research Unit, and
Baltimore VA Center Stroke of Excellence

University of Maryland School of Medicine
Division of Gerontology
Baltimore VA Medical Center
10 North Greene Street
GRECC (BT/18/GR)
Baltimore, MD 21201-1524

(Phone) 410-605-7119
(Fax) 410-605-7913 (Please call phone number above prior to faxing)
jsorkin at grecc.umaryland.edu

>>> "hadley wickham" <h.wickham at gmail.com> 11/24/07 10:56 AM >>>
> What do you mean here?  Surely all packages authors aim to provide
reliable
> and effective software.  If they know that they are offering
something
> unstable, they should say so clearly.  In fact, they should wait
until it is
> stable.  Most R users are not researchers, but users.

Now I'm beginning to wonder if you have ever used R ;)  Yes, they
should wait until it is stable and free of bugs, but does this really
happen in reality?  The quality of packages varies widely, and even in
the best package it's difficult to find all bugs before release?

> This web reference shows that you are thinking of a quite different
type of
> review.  What help would that be to a user?  (However helpful it
could be to
> another researcher).

I was suggesting that data would be useful when selecting which
packages to review, not as part of the review.

> You're right, the thread has moved on.  No one would read either
1000
> reviews or 1000 brief paragraphs.  Reviewing should help to raise
standards.
>  Good reviewers would point out connections with other packages and
make
> comparisons.  (Which does take us partway back to the original
thread.)

Which moves somewhat back towards my original suggestion of review
articles.  To me, an article which compared and contrasted four or
five packages on a given topic would be much more useful than an
article which reviewed only a single package.  I think basing reviews
around a specific topic/methodology would be more useful than basing
them around a single package.

Hadley

-- 
http://had.co.nz/ 

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help 
PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html 
and provide commented, minimal, self-contained, reproducible code.

Confidentiality Statement:
This email message, including any attachments, is for th...{{dropped:6}}