[IAEP] activities.sugarlabs.org

[Greg Dekoenigsberg]

On Wed, 18 Jun 2008, David Farning wrote:

> On Wed, 2008-06-18 at 10:44 +0545, Bernie Innocenti wrote:
>> David Farning wrote:
>>> It looks like my top priority will be converting
>>> https://addons.mozilla.org/en-US/firefox/ to work with sugar and it's
>>> ecosystem of activities.
>>
>> Hello from Kathmandu!  You will have all my support on it.
>> Ideally, I'd like activity authors to be able to host their projects
>> on sugarlabs.org and manage it easily and autonomously.
>>
>> We could evaluate one of the many forks of Sourceforge, or something
>> lighter weight.
>>
> Addons.mozilla.org also known as amo looks pretty promising as a 'one
> stop' location for activities.
>
> In terms of work flow the process for a developer is:
> 1. Activity authors register their activity with amo.
> 2. Author uploads latest release into sandbox.
> 3. Reviewer verifies new upload works correctly and publishes the
> upload.
> 4. Return to 2.
>
> All of this happens through a nice set of web interfaces.

There's also localization flow.  How well it works, I'm unsure.

> We have all seen how amo works from an end users perspective.
>
> cons:
> 	The amo code is pretty dense.
> 	The amo code is quite hard to read because of all the efforts made
> towards serving addons efficiently.

These are the cons that I ran into as well.  Another con is the speed with 
which the codebase changes.

> pros:
> 	From amo's point of view the work flow between mozilla updates and
> sugar activities is exactly the same.
> 	Very fast, secure and scalable.

These were also my biggest pros.  Another: an actual community developing 
the code, with whom you can consult when you run into trouble.  That 
should be a huge win.

> Progress:
>  Done:
> 	Got amo running on home server.
> 	Studied work flow from end user and developer point of view.
>  In progress:
> 	Studying database schema.

I got a bit farther in my efforts, but unfortunately, my work bitrotted 
over the last 18 months since I first made the effort, and the AMO 
codebase has changed quite a bit.

There's also a large question that needs to be answered: will the 
necessary changes be carried upstream, or will they constitute a fork?

Are you working with anyone at AMO on this problem?  I'm guessing you know 
some people, but if you don't, I can comb through my notes and maybe put 
you in touch with some other people.

Well done on tackling this problem, David, and good luck.  Maybe I'll jump 
in and help at some point... or maybe I'll just cheer wildly from the 
sidelines.  :)

--g