I want to address Isarra's feedback (below), but before let me complete S's
reply to Brian.
On Mon, Jul 6, 2015 at 7:41 AM, Brian Wolff <bawolff(a)gmail.com> wrote:
In essence, this is an advertisement aimed at
programmers not
associated with the Wikimedia movement, to use various Wikimedia APIs
in their programming projects that are unrelated to Wikimedia. In a
sense, targeting these programmers as a distinct group of users, and
trying to convince them to use our "product".
Is that right? If it isn't, may I suggest that this should be the
intended purpose?
This is the intended purpose, yes. The rest of your reply is spot on.
Comments:
If so, I think the content of the developer hub should have a little
bit of a different focus. It should concentrate on
things that make us
unique, and things that are likely to have wide applicability and
value outside of Wikimedia.
Yes. We have been underestimating the importance of looking at our APIs
with external eyes. We are so used to look at our own use cases, and so do
most of the developers S relies on to document APIs. Proposals for projects
/ people we should contact and interview to gather ideas are welcome.
(Also, the name is very confusing.
After more than a year we haven't solved the problem of finding a clean
association with "developer" for the many reasons many people have
provided. It has been proven difficult to add the "data" aspect too. The
feedback received after the announcement of the prototype confirms this
problem.
As a solution, I'm proposing to focus on the Api aspect, improving the Api:
namespace in
mediawiki..org and finding a project name that stresses the
Api aspect. Once the discussion about the namespace at
https://www.mediawiki.org/wiki/Thread:Project:Current_issues/Adding_a_dev_n…
is resolved, we will rename the project accordingly.
The very fact that in
[[mw:dev.wikimedia.org]], it states its not for
MediaWiki "hackers"
despite the fact that in our community (And particularly in the larger
Wikimedia community), "developer" as a word is much more commonly used
than the word "hacker" is, for what the document means by "hacker",
should attest to the confusingness of the nomenclature).
Yes, yes, edited.
In particular the proposed persona's suggest a different target than
what I commented above.
Also agreed, and as soon as the namespace discussion is clear, we will
update those personas accordingly. The more we got into details, the
clearer it has been that we should narrow the focus of this project. Let's
produce something interesting and useful for those third party developers
that could use our web Apis in their own projects. We also need to make
happy the other personas, but one step at a time.
On Thu, Jul 2, 2015 at 9:15 PM, Isarra Yos <zhorishna(a)gmail.com> wrote:
It should perhaps be noted that this seems a
continuation on a general
trend to avoid learning how to work with the communities and existing
designs. It may be faster and less frustrating to simply sod off and do
something new somewhere else, but this does not solve the existing problems
- all it does is introduce more inconsistencies and more things that need
to be consolidated later, if they survive at all. Except they usually
don't, because ignoring what people do and use and expect does not lead to
practical designs, it leads to things people don't use, maintain, or
contribute to, which brings us back in part to the technical debt
associated with having so many different wikis.
But when it comes to working with what already exists, I know from
experience that this is far from easy, and in fact often incredibly
frustrating in practice. Thing is, though, we have to - the existing
products need to be worked with in order to move forward. That's just how
it is.
As we tend to say about the unsolicited redesigns of (en) wikipedia, it's
always a lot easier when you don't have to design with reality as a
constraint. There's a reason we never use any of them.
The problem I have with this reasoning is that we are not proposing a
different skin for Wikipedia. The design of this site doesn't need to be
efficient for the multiple use cases and the heavy legacy of Wikipedia.
There is more than Vector out there. We can allow ourselves to experiment
with something fresh, and then fine tune or revert.
Vector is the skin powering Wikipedia and hundreds of Wikimedia sites (and
more), and for this reason changing anything there takes a lot of effort
discussing and implementing (see Winter, and see several other projects
that took so much time and brought so little changes to our users). We
don't have design resources for this project, but we don't want to renounce
to bring a fresh look to it. We are recycling a skin that already existed,
we are filing bugs, and then Prateek, S, and maybe others contribute
whenever they have time, motivated by their willingness to see that fresh
look arriving to users.
It is not strange to find developer sites not following the user experience
of their super-popular products. This gives them freedom from the "product
people" and this allows them to focus on the specific purpose and audience
of their developer sites. Having a skin deployed in
mediawiki.org that
looks contemporary and not-like-Vector, not-like-Wikipedia cannot be
harmful. As said, the worse that can happen is that we must revert and
fallback to Vector. Fine, at least we can say that we tried.
--
Quim Gil
Engineering Community Manager @ Wikimedia Foundation
http://www.mediawiki.org/wiki/User:Qgil