[wp-hackers] GSoC Proposal: JSON REST API

[Mike Schinkel]

On Apr 12, 2013, at 4:23 AM, Bryan Petty <bryan at ibaku.net> wrote:
> Which is why it's not even being considered for merge to trunk during
> (or for quite some time after) GSoC.

I will repeat: Implementing a JSON API as a GSOC project is a BAD idea.  

To constrain the architecture decisions to 2 months minus the time required to fully implement is a BAD idea.  To effectively have one person with one mentor choose the architecture for a JSON API that could affect 70 millions sites is a BAD idea.

A GSOC JSON API will be treated as the defacto-standard JSON API implementation planned for WordPress core and thus more than a few people will use it. Implemented API code has great inertia. Architecture choices will quickly be "baked in" from usage and the cry to halt change in the API will be great, especially is someone influential with core builds to this API.

A JSON API has attributes that other GSOC projects will not have.  APIs are different from UX; people can adapt to a changing UX[1]; programs written to API cannot adapt unless someone devotes time or funds to modify them. 

Please do not do this; make it a Trac ticket instead. Organize a KickStarter campaign if it's about the money; I'll back it.

-Mike
P.S. Again, this is not personal. I have great respect for both you Bryan and Ryan and I hope my objection to this does not sour you on me personally. This is because I truly believe a JSON API is too important for the GSOC process.

[1] http://wordpress.org/extend/plugins/mp6/