Re: [gentoo-soc] Re: [GSOC 2012] Recruiting Webapp idea

[Petteri Räty]

On 05.04.2012 23:08, Gaurish Sharma wrote:
> Hi,
> Kindly read the replies inline:
>

Where else? :) In Gentoo mailing lists every should be responding inline.

>
>
> On Fri, Apr 6, 2012 at 12:50 AM, Petteri Räty <betelgeuse@g.o> wrote:
>> On 04.04.2012 21:18, Gaurish Sharma wrote:
>>
>>
>>> b) the User interface is old, dated & just boring. We need a fresh theme.
>>
>> CSS work is nice as long as we don't spend too much time on it.
>> Usability is more important than looking good. In some situations they
>> of course go hand in hand.
> -->
>
> I suggest that we switch to responsive design based on twitter
> bootstrap[1], so that webapp is accessible from all mobile,tablets &
> desktop browsers[1]. The toolkit has pretty good styling & is usable
> by default, however. I would customize according our needs.
>

twitter bootstrap is a nice toolkit. It's nice if we get a working
mobile interface on the side but our primary target is ordinary desktop.
I don't think people want to fill their quizzes with phones. Of course
we can do a poll as I am just using my unverified assumptions here.

>> The hobo plugin we used currently only has a stable release out for
>> rails 3.0. There does seem to be work going for 3.1 so we could target
>> that instead after checking the hobo timetable for getting 1.4 out
>> (which is needed for rails 3.1).
> --->
>
> v3.0 vs v3.1 -- There are a few backwards-incompatible changes in
> Rails 3.1 which affect the future upgrade path if we choose to develop
> in 3.0 only. rincipal among these are a new convention for the
> placement of images, stylesheets, and JavaScript files—collectively
> known as assets—and a change in the default JavaScript library from
> Prototype to jQuery. Further, 3.1 has cool features like Asset
> pipeline,Reversible migrations which we would miss out we stick with
> 3.0.
>

You can write reversible migrations with the current version already.
The new feature in rails 3.1 is that it can automatically deduce the
down migration from the up migration in most cases.

> However,
> With every new Rails release, there is a time delay while the rest of
> the ecosystem catches up. Now, hebo severely lagging behind. so I
> propose, we dump hobo & we writte the app as per the requirement while
> importing existing data on latest version of rails.
>
> The reason I want to dump hobo is:
> 1) it is lagging behind in support for latest releases

The lag in getting new releases is certainly annoying.

> 2) It is not core rails.

Why is this a problem? Most rails projects end up using other gems than
just rails. Hobo of course is more invasive than than most plugins so
the learning curve must be weighed against potential benefits.

> 3) Hobo is good for protying but for apps that are required to go into
> production like recruiting app, traditional development model is
> preffered
>

There are multiple hobo applications in production. I agree that Hobo is
not perfect. We can migrate away to plain rails but my hunch is that it
would take too much time for too little benefit.

> I hope this makes sense.
>
>
>
>
>> You should scope the draft proposal so that you yourself think there's
>> enough work to fill the whole GSoC for you. If think the items listed so
>> far are not enough for that feel free to ask for more ideas.
> -->
>
> Sure, please let know what more than I can do
>

Here's a couple things I can think of:
- integrate with Gentoo LDAP. I think there was a project last year that
was allowed to do that so we can probably too.
- machinery for comparing answers for copies
- meeting bot integration (similarly to the council web app from last year)

These are all of lesser priority than what we have in bugzilla and UX.

Regards,
Petteri