Time |
S |
Nick |
Message |
00:44 |
|
pdurbin |
nice |
08:28 |
|
|
philbot joined #sourcefu |
08:28 |
|
|
Topic for #sourcefu is now http://sourcefu.com | logs at http://irclog.greptilian.com/sourcefu/today |
09:05 |
|
|
semiosis joined #sourcefu |
09:56 |
|
|
tumdedum_ joined #sourcefu |
09:59 |
|
|
tumdedum_ joined #sourcefu |
16:56 |
|
|
tumdedum_ joined #sourcefu |
19:56 |
|
pdurbin |
Hmm. Toward the end of https://twit.tv/shows/floss-weekly/episodes/383 at 50:40 or so Randal Schwartz starts talking about how large the CPAN is. "It seems like there's a lot more work to go through if you're going to be in one of those other ecosystems." However, npm (and others) are much bigger according to http://www.modulecounts.com which is powered by https://github.com/edebill/modulecounts |
19:57 |
|
pdurbin |
Interestingly, he goes on to say that CPAN gets 60-70 uploads per day but http://www.modulecounts.com only shows 4 per day for CPAN. (And 667 per day for npm.) I'm not sure who's right. |
20:00 |
|
pdurbin |
sivoais: any thoughts on this? |
20:02 |
|
sivoais |
it's definitely not 4... maybe around 30 on average as can be seen at <https://metacpan.org/recent> |
20:02 |
|
sivoais |
*cough* the last few Alien::* modules on there are me :-P |
20:04 |
|
sivoais |
wait, is this new uploads? |
20:04 |
|
sivoais |
as in, first release |
20:06 |
|
pdurbin |
sivoais: not sure. I was thinking an issue could be filed at https://github.com/edebill/modulecounts/issues if some calculations are not being done properly. |
20:14 |
|
sivoais |
I'd have to look more closely later. |
20:15 |
|
sivoais |
the code structures is a bit odd. I wouldn't expect the regex that do the extraction to be in the db/ directory |
20:20 |
|
sivoais |
I still think module counts don't tell me anything unless I download and look at what each thing does |
20:21 |
|
sivoais |
that's where data mining would be really useful |
20:21 |
|
sivoais |
It'd be nice to get clusters a bit like <http://mapofcpan.org/> |
21:33 |
|
pdurbin |
sivoais: but CPAN isn't the biggest anymore, is it? |
21:35 |
|
sivoais |
no, it doesn't look like it |
21:41 |
|
sivoais |
but I can say that MetaCPAN has the best website :-P |
21:41 |
|
sivoais |
the only one that comes close is Hackage for Haskell |
22:11 |
|
pdurbin |
:) |
22:12 |
|
pdurbin |
well, that's the main feedback I'd want to give to Randal... that what he's saying about CPAN was once true but no longer |
22:13 |
|
pdurbin |
from what I can tell he's on a cruise right now... probably not the best time to ping him :) |
22:18 |
|
sivoais |
I still want to look at the modules with text mining and graph theory. I think there will be some interesting things just looking at how different modules use their dependencies. |
22:18 |
|
pdurbin |
yeah. potentially a very interesting dataset |
22:23 |
|
sivoais |
I haven't used this tool yet, but it looks like it could help people that manage large FOSS projects see how people work together |
22:23 |
|
sivoais |
<http://siemens.github.io/codeface/> |
22:24 |
|
sivoais |
very cool bit on the examples page |
22:24 |
|
sivoais |
<http://siemens.github.io/codeface/img/openssl_9.dot.png> |
22:24 |
|
sivoais |
shows how different people in OpenSSL contribute to specific components of the project |
22:25 |
|
pdurbin |
huh, yeah, neat graph |
22:25 |
|
sivoais |
The data set comes from the mailing list. It looks like the words are stemmed. |
22:26 |
|
pdurbin |
I think https://www.openhub.net tries to capture some of this |
22:28 |
|
sivoais |
hah, I started trying to put all my projects on there, but it took a long time |
22:39 |
|
pdurbin |
sivoais: did you get any value out of the data after you put some projects there? |
22:47 |
|
sivoais |
pdurbin: I think the main value for me was being able to see when I was working on a given project. They have a good overview page for the user. |
22:49 |
|
sivoais |
and when you go to the individual project page, showing the COCOMO model numbers ( "X years of effort" ) isn't directly useful for me |
22:49 |
|
sivoais |
but it does serve as a way to motivate more development |
22:50 |
|
sivoais |
and at the same time... a little bit of justified shaming :-P |
22:50 |
|
sivoais |
when it says "very low number of source code comments" |
23:23 |
|
pdurbin |
heh |