Log file Viewer

[07:11:12] <[Tritium]> about how many projects are on pypi?

[07:14:30] <tdsmith> https://pypi.python.org/pypi says 81,211

[07:14:46] <[Tritium]> ...i didnt realize it was on the front page

[07:15:16] <[Tritium]> well that blows the old perl argument out of the water

[07:16:10] <[Tritium]> If you want a language that has a large library database available, the choice is easy: node.

[07:16:55] <tdsmith> otoh https://github.com/const-io

[07:21:16] <[Tritium]> so this site exists http://www.modulecounts.com/

[09:57:12] <dstufft> [Tritium]: tdsmith I mean, directly comparing counts between different languages is a bit apples-to-oranges since each language has a different culture around how you divide up the work as well as a different standard library-- Node for instance probably doesn't have as many things implement async IO as Python does, since Node has that built in, but Node's ecosystem values splitting things up into smaller packages over using larger ones

[09:57:26] <dstufft> Unfortunately, there isn't a great metric to compare them :[

[09:59:55] <[Tritium]> dstufft: rate of package update

[10:03:27] <ionelmc> there was this 3d graph of all the pypi pages, anyone has the link?

[11:37:42] <xafer> dstufft: regarding the download dataset you provided on distutils-sig, is it only the effective downloads or does it also count the cache hit ? (Typically when pip decides not to redownload a sdist)

[12:20:47] <Wooble> [Tritium]: would it help Python adoption if I released 10 million 2-line modules?

[12:30:13] <[Tritium]> Wooble: if you released 10 million 2 line modules, you would end up banned from pypi and the packages would be removed

[12:32:20] <Wooble> but what if they're all essential useful functions like, uh, adding a few spaces to the left of something? :)

[15:10:57] <dstufft> xafer: it's only things that actually hit Fastly (a 200 or 304 response), if pip decides not to download the file because it was cached locally, then we can't see that

[17:45:38] <ich_dien> Hello all