I'm just now looping back to this and apologize for the delay, and it'll
probably be another several days before I check this thread again.
My current thinking is that the portal is more of a content page than
technical infrastructure. For that reason, the decision-makers about the
content of that page should be the community, not WMF. Perhaps there has
been a community discussion which specifies the scope of decisions that the
community is willing to delegate to WMF, but I'm unaware of that. I think
that such a discussion should happen before WMF effectively takes control
of the page content. Otherwise, to me this looks like scope creep by WMF
over a content page.
It sounds like WMF is collaborating with community members regarding page
content, which is good. It seems to me that there should also be some
specifications about which decisions get made by WMF and which get made by
the community. Otherwise, this ambiguity is just asking for trouble.
On Tue, May 24, 2016 at 6:43 PM, Dan Garry <dgarry(a)wikimedia.org> wrote:
On 24 May 2016 at 17:01, Pine W <wiki.pine(a)gmail.com> wrote:
Perhaps someone can enlighten me: did WMF consult the community about its
wish to change the content on this public-facing content portal? (A portal
survey does not carry the same weight as a community RfC). If so, when,
where and how was there a public discussion with proportionate notice and
scope, and who closed the public discussions?
No RfC was conducted. These changes were made on the basis of an A/B test
which showed that they likely improved the ability of users to navigate to
sister projects. By performing tests, we can speak with confidence about
what effect our changes have. A report was published
the test; I'd encourage you to read it and give us any comments you have.
The Discovery Department has no plans to perform consultations for every
change we make. We use testing and data to guide our decisions. With large
changes, we may perform consultations to also guide us, but this was not a
large change at all. As the report linked above notes, very few users
interact with these UI elements on this page; of the 22,500 recorded
sessions in the test, only 45 clicks on the sister project links were
registered. This was enough for the statistical test, but it shows you the
magnitude of the engagement here.
The portal repository continues to have volunteer contributions to it,
although there are fewer now as the team handles many things such as
updating the statistics. If volunteers are daunted by the prospect of using
git and gerrit, Discovery would be happy to offer training so that they can
continue to contribute.
Lead Product Manager, Discovery
discovery mailing list