[wp-hackers] GSoC 2011

[Otto]

On Wed, Mar 30, 2011 at 10:20 PM, Thomas Van Eyck
<thomas.vaneyck at gmail.com> wrote:
> This discussion is getting of track. The do's and dont's of DB design are
> important, but not the key topic of the discussion. It has been interesting
> though :-)

I think if you're wanting to improve the media system, that's perfect
for a gsoc project, but perhaps you should focus more on the interface
and the end goals instead of how you're wanting to get there from
here.

The media handling system is quite robust, but it's the interface that
people really see. If you had some good goals for what you want people
to see, or how an organization system should work from the user's
perspective, then you'd have a good project.

It's also important to recognize that the system isn't "legacy", as
such. Attachments are stored in the posts table for a reason. That's a
design feature, not a legacy architecture. The DB design has evolved
over time to be highly fast and flexible. There's lots and lots (and
friggin' lots) of code dealing with it in specific ways, and generally
speaking you can fit pretty much anything you want into it. It's just
a matter of rethinking. And I say this as somebody who has had to
force myself to unlearn "third normal form" in learning scalable
web-database programming. ;)

So focus more on that end result. Maybe think about some mockup
screens, or what sort of functionality in the interface would be
useful to organize media. Heck, just adding an organization system for
media would be a big project all by itself.

-Otto