Brian Wolff wrote:
Maybe the grant includes funds for hiring code review
resources (ie
non-wmf people with +2. We exist!).
For what it's worth, you're exactly the type of person I would like to have
working at the Wikimedia Foundation. I love your posts here; thank you for
taking the time to write them.
Figuring out what level of technical support we can give to non-Wikimedia
Foundation projects is a really important issue, in my opinion.
Brian Wolff wrote (in a related thread):
Ostensibly this is done in the name of:
Any technical components must be standalone or
completed on-wiki.
Projects are completed without assistance or review from WMF
engineering, so MediaWiki Extensions or software features requiring code
review and integration cannot be funded. On-wiki tech work (templates,
user scripts, gadgets) and completely standalone applications without a
hosting dependency are allowed.
Which on one hand is understandable. WMF-tech has its own priorities,
and can't spend all its time babysitting whatever random ideas get
funded. So I understand the fear that brought this about. On the other
hand it is silly, since a grant to existing tech contributors is going
to have much less review burden than gsoc/opw, and many projects might
have minimal review burden, especially because most review could
perhaps be done by non-wmf employees with +2, requiring only a final
security/performance sign off. In fact, we do already provide very
limited review to whatever randoms submit code to us over the internet
(regardless of how they are funded, or lack thereof).
Erik seems to be pushing toward a model that favors using OAuth and the
MediaWiki API over "deep integration" that comes with a MediaWiki
extension. He recently mentioned this here:
https://lists.wikimedia.org/pipermail/glamtools/2015-February/000343.html
He may be right that development for deployment to the Wikimedia
Foundation cluster may not be the best approach for every project, but I
think this view overlooks all the very real benefits that extension
deployment includes. There's a documented process that has safety checks
such as putting the code in Gerrit and having a security review. Checklist:
<https://www.mediawiki.org/wiki/Review_queue#Checklist>. Process:
<https://www.mediawiki.org/wiki/Writing_an_extension_for_deployment>.
MediaWiki is the platform. Features include persistent database or file
storage, user authentication, internationalization, a usable Web API and
user interface, and more!
If IEG grants were allowed in this area, it would be
something that the
grantee would have to plan and account for, with the understanding that
nobody is going to provide a team of WMF developers to make someone
else's grant happen.
Yeah, my understanding is that Sue was behind this hard rule and times
have changed. I guess this would be a matter of Siko and her team
re-petitioning Damon, Erik, or Lila to soften this rule, probably by
appending a "or have a detailed code review plan in place with appropriate
sign-off/endorsement" clause. This code review plan would be some kind of
template where people can do due diligence to try to ensure that their
code review needs will be met.
More broadly, in terms of getting code deployed to the Wikimedia
Foundation server cluster, we have at least three major code review areas:
security, performance, and architecture. A code review plan (for grants
and non-grants alike, to be honest) that addresses at least these three
areas, plus user acceptance, as you mention, would be fantastic, I think.
And/or we can explore the model proposed by dan entous:
---
instead of having to write a grant requests and/or seeking other forms of
funding, establish a grant or funding committee that looks for projects
and developers that have proven helpful and have added value to the
community. then award them with funding without them having to ask for it.
---
Politically, I think its dangerous how WMF seems to
more and more
become the only stakeholder in MediaWiki development (Not that there
is anything wrong with the WMF, I just don't like there being only 1
stakeholder).
Yup. Other groups such as Wikimedia Chapters are also interested, but all
most of the funding streams go through the Wikimedia Foundation for
redistribution at this point, as I understand it. Maybe a MediaWiki
Foundation still makes sense... Brion and others have been pushing for a
wiki hosting platform (that isn't the ad-plagued Wikia, heh):
https://lists.wikimedia.org/pipermail/wikitech-l/2015-January/080171.html
MZMcBride