Re: OPW; Where does the 500$ for each GSoC goes?

[Sébastien Wilmet <swilmet gnome org>]

On Tue, Sep 16, 2014 at 01:27:47PM +0100, Emmanuele Bassi wrote:
> On 16 September 2014 12:01, Sébastien Wilmet <swilmet gnome org> wrote:
> > I know that with a good documentation (i.e. a book) and good development
> > tools, developers will be more attracted by the Gnome platform.
>
> that's not nearly enough.

It's maybe not enough, but it's needed too in my opinion.

> the existence of a book during GNOME 1.x hasn't really changed much in
> our DX, and books for free software platforms tend to go out of date
> by the time they hit the printer.

GLib, GObject and GIO are pretty stable. If you read the GLib section in
GGAD, you'll see that it's still relevant, just a minor update is
needed.

For GTK+, Benjamin explained last year that the plan is to have GTK+ 4
as stable as GTK+ 2. Indeed GTK+ 3 is quite unstable, so writing lots of
details about GTK+ will probably be quickly outdated (API deprecations
is another problem).

> > There are important areas where Gnome is lacking. If the Gnome
> > Foundation can give a little more financial support to improve those
> > weak areas, it would be great.
>
> please, enumerate them, and have actionable deliverables to verify
> that they are, in effect, improving weak areas.

- Improving the developer experience. Christian Hergert is working on
  Builder and he will soon start a fundraising campaign. This is an
actionable item.

- Writing a book, another actionable item.

- When a company is interested by the Gnome platform, help them. Philip
  said for example:

<pvanhoof> My issue with gnome is that it wants to hate trying to help the
           (european) commercial world
<pvanhoof> gtk+ didn't really help the maemo-gtk+ stuff
<pvanhoof> That as a result meant that meego-harmattan got qt based
<pvanhoof> That's history, of course. But we did try hard to convince people at
           redhat like mclasen of the necessity to cooperate with nokia back then
<pvanhoof> Yes we did

When such collaboration is needed and is not done, it is another
actionable deliverable.

> I'd contend that maintainers have already one job to do, and they are
> not really good at writing documentation for beginners.

GGAD is well written, and was written by a core GTK+/Gnome developer.

A good scientist should be able to explain his or her work to a young
child, there was a citation about that.

> nevertheless, we do have core developers writing documentation — in
> the form of wiki pages, like the HowDoI documents, or the API
> reference examples.

A book would have the advantage that it's a single source of
documentation for a beginner. Currently the documentation is cluttered:
a beginner needs to read this there, that there, in the API reference,
in the wiki, in another tutorial found on the web, etc.

> > Like Havoc Pennington describes it [2], GGAD
> > is an "advanced internals"/"conceptual understanding" kind of thing. And
> > this kind of book is needed to get more experienced developers,
> > beginners would learn more quickly, and better software would be
> > created. For the long-term it would be beneficial. It would be a good
> > investment.
>
> you're skimming over various fundamental bits, and making some
> unconnected leaps, like: a beginner's documentation does not look at
> all like an expert documentation; you cannot improve the former just
> by having the latter. the ability for a beginner to learn "more
> quickly" is not a result of a lack of low level documentation, and
> there's really no proven connection there.

Ideally the book should contain both documentation for beginners to get
started, and advanced chapters for having a better understanding.

The goal of the outreach programs is to have more experienced
developers. If there is no good material to become quickly an
experienced developer, the software written will have a bad quality. Of
course the mentor should guide the student so that he or she learns
quickly what is needed, but a book would be much simpler, and many other
people could learn the Gnome platform more quickly.

For example, in GGAD the GTK+ Basics chapter contains an explanation of
size requisition and allocation, widget life cycle, etc. The advanced
chapters talk about GDK and writing GTK+ widgets. If you want more
experienced developers, this kind of information is needed, and with a
good book, more people will learn that, more people will better
appreciate GLib/GTK+ and stick with it instead of switching to Qt.

--
Sébastien