This is a long thread that I just caught wind of, but I thought I'd interject with a few notes.
The E3 team is going to start doing experiments on the account creation process, starting with the signup page. The front-end is going to be reworked to conform to the design team's new "Agora" standards, like so:
http://www.mediawiki.org/wiki/Account_creation_user_experience
As for the API: we are going to make some improvements to it as part of our first experiment. Yes, it is functional now, but we need to do a bit of hacking to support our proposed UX improvements, as well as make it more consistent with WMF JavaScript guidelines. If someone wants to improve the signup template in core, that would be excellent - but since our team needs to move quickly, we're likely going to just going to make our own fork of the extension and display the proposed template for users in a small experimental bucket.
Feel free to email me if you have questions. :-)
Cheers, Munaf
On Fri, Aug 17, 2012 at 8:30 AM, Tyler Romeo tylerromeo@gmail.com wrote:
With items like SpecialPages and API classes, it is indeed a possibility to first make it an extension and then integrate it, primarily because there is little difference in how a core SpecialPage/APIBase is implemented and how an extension is implemented. However, this workflow is not required.
The problem is that the LoginForm class is old and run-down, and we shouldn't really be basing code off of it. A better way to go about it would be to make use of the newer MW infrastructures like FormSpecialPage and Status to make a cleaner implementation. The account creation API currently in Gerrit is actually a bit of a hack (as is the Login API and many other similar modules) because of the fact that there is not a good separation between application logic and UI in many core features of MW. It would be much preferred to fix this then to pile on top of the current way things are implemented.
*--* *Tyler Romeo* Stevens Institute of Technology, Class of 2015 Major in Computer Science www.whizkidztech.com | tylerromeo@gmail.com
On Fri, Aug 17, 2012 at 11:09 AM, Akshay Agarwal < akshay.leadindia@gmail.com
wrote:
I am not sure how stuff works now but at the time I was working on this project, any features missing in the MediaWiki codebase were first developed as an extension, reviewed, tested and then integrated in the core.
The current version of the SpecialPage looks similar to LoginForm because it was indeed derived from there & one of the main goals for this project was to remove the account creation code from SpecialUserLogin & put it inside its own SpecialPage. I do realize that some refactoring might
still
be needed but I definitely can say that the efforts to do that would be much lesser than rewriting the entire module because the new module would again have to go through similar iterations through which SignupAPI
already
went.
Tyler, I really appreciate your efforts in developing a new API & I would encourage you to contribute with all the learning that you have had in
this
project to getting SignupAPI deployed because the project involves many more things than just an additional API, many of which I have conveyed in my previous mail.
On Fri, Aug 17, 2012 at 8:27 PM, Derric Atzrott < datzrott@alizeepathology.com> wrote:
User account creation is something that is pretty critical to the
MediaWiki
infrastructure. If we're going to be completely revamping the signup
page,
it
should not be done in an extension.
Furthermore, looking at the extension's code now, there are numerous
design
problems that would need to be fixed if this were to be deployed
(primarily the
fact that the entire special page looks pretty much like a copy of
LoginForm's
account creation interface).
Not to mention that an account creation API is something that the core
needs and
site admins should not have to rely on an extension to install it.
I'm welcome to a rewrite and refactoring of the LoginForm, if that's a
goal we
want to aim towards.
Still we should take note of the lessons he learned when he made his extension and apply them to development of a API in the core, should we
go
that route.
I agree that this should be a core feature and not an extension. This seems like the sort of thing that many Wikis will need and I'm somewhat
surprised
the request for someone to make it doesn't come up more often.
Thank you, Derric Atzrott
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l