|
11:30
Nemokosch joined
|
|||
| Nemokosch | Hello. This has been bugging me for a while... "Popular Dists" doesn't look good, content-wise. First, "zef" and "cro" are simply duplicated. Second, and this is probably a more fundamental problem: it's full of legacy stuff. I suppose when the ecosystem was pretty much github itself, giving stars to repositories was much more common. | 11:33 | |
| It's pretty bad marketing that under the name "popular dists", one sees mostly abandoned old stuff. If there is no obvious good way to improve it, I would be all in for removing that part altogether. | 11:35 | ||
| lizmat | I agree to an extent: the problem with going by number of stars, you by definition favour longer existing modules | 11:36 | |
| also, with people installing through zef, they're mostly unaware that there is something like a github repo at which they can show their appreciation | |||
| and you need a Github/lab login for that anyway, which many users may not have | 11:37 | ||
| in any case, at least the Popular Dists should pay attention to the underlying repo, and filter out identities that map to the same repo | 11:40 | ||
| and a similar check should be done for authors, filtering out identities that map to the same repo | 11:41 | ||
| I guess someone should make a PR :-) | 11:42 | ||
| Nemokosch | "the more stars on github, the more popular the distribution" isn't a bad approach in theory, however, the outcome does matter. Anything that directs you to Bailador, Alma, panda and even JSON::Tiny to some extent, needs some change. Even showing random distributions by default would be better marketing. | 11:44 | |
| lizmat | perhaps the popular dists should be replaced by "highest number of downloads" | 11:45 | |
| this would favour the distributions on zef, which by definition would be more modern | |||
| Nemokosch | That may be a better idea | ||
| lizmat | as the CPAN / p6c downloads are unknown | 11:46 | |
| and since we want to phase out CPAN and p6c as backends, would be an extra incentive for people to move their modules to fz | |||
| *fez | |||
| Nemokosch | of course that's also a compromise but imo a better compromise than the current one, which is mostly biased towards p6c era stuff | 11:47 | |
| lizmat | which, almost by definition, is outdated, bordering on obsolete | ||
| Nemokosch | yep | 12:02 | |
|
12:13
Nemokosch left
13:23
jgaz joined
|
|||