This is a question that maybe the rOpenSci community already figure out the answer, but what if someone prefers GitLab or Bitbucket to host their package and still wants the PyOpenSci badge/approval/review? Could we review those and give the same status? Or maybe a different status like a badge “reviewed by PyOpenSci?”
This case may also happen with packages that even though are on GitHub but cannot be transferred to the PyOpenSci organization. Maybe a fork in the PyOpenSci (that is kind of a bad idea IMO b/c forks go out of sync quickly)? Or the same badge as above?
The other option is, of course, to just reject those packages by default, making the transfer mandatory for the approval. I have to confess that I am on the fence on this one. Every time I think about it I come to a different conclusion. More work vs simplicity, curation in a single repo vs external badges, etc.
Gosh - another great question @ocefpaf
i’d like to ping a few people here including @mbjoseph and @noamross
It is my understanding that ropensci encourages people to transfer to the ropensci org but does NOT require it. I just checked our dev_guide and didn’t actually see this discussed in detail. I would think that some people have good traffic to their repo already and might not want to move to github.i think i’d like to support that but would love to hear other thoughts.
What has worked for ropensci?? We should definitely update our guides once we decide!!
I haven’t used gitlab a lot but don’t see why it would be an issue. so open to feedback here.
Also – i think we should put a list of packages that are in our POS org on the website like what ropensci already has
Noam - do you guys have a database to support this list? it’s really nice.
UPDATE: i asked the pyOpenSci folks and they did say that they require a git based repo which would make something like gitlab or bitbucket ok. The transfer will always be optional so we can keep a list of packages on our website that we’ve accepted and that list could link to different home repos be it github, gitlab or bitbucked.
I’m really glad that this was the decision (only saw this discussion now). Personally, I would hesitate to send some of my software for review if it was mandatory.