On Tue, Aug 05, 2003 at 12:38:29AM -0400, Steve Grazzini wrote: > If it matters, CPAN gives (or cites, really) good reasons why > those statistics shouldn't be provided. > > FAQ: > > Does CPAN provide download statistics for authors? > > No we don't. http://xxx.lanl.gov/help/faq/statfaq sums up our > thoughts on the matter quite well.
from xxx.lanl.gov: > It could be argued perhaps correctly that statistics may provide some useful > information at least on the relative popularity of submissions, since the > distributed access and other factors may be subsumable into some overall scale > factor. But even this information is ambiguous in many cases, and publicizing, > even when accurate, could merely accentuate faddishness in fields already > excessively faddish. Ambiguous, but I think CPAN stats could be potentially useful. Just as Schwern's Kwalitee is an approximation of quality, and does not aspire to providing "perfect" indicators of quality, a metric on CPAN useage could provide some sort of picture of the Perl community at large, as long as they include the caveats. Perhaps we need some Statisticks instead of Statistics :) > Most significantly, however, there is a strong philosophic reason for not > publicizing (or even saving) these statistics. When one browses in a library > it is very important (in fact legislated) that big brother is not watching > through a camera mounted on the wall; for the benefit of readers it is very > important to maintain in every way possible this sense of freedom from > monitoring in the electronic realm. This also is true, however legislation prevents making an *individuals* browsing patterns public. Any good library closely monitors it's circulation statistics and looks for patterns to guide future collection development. So if titles in an area such as Camel Biology are continuously checked out by a diverse group of people, the Library can decide to better serve its community by purchasing more books in this area. //Ed
