Re: Certification for GNOME apps

[Alan Horkan <horkana maths tcd ie>]

On Thu, 14 Jul 2005, [utf-8] Danilo Šegan wrote:

> Date: Thu, 14 Jul 2005 12:06:56 +0200
> From: "[utf-8] Danilo Šegan" <danilo gnome org>
> To: Andrew Sobala <aes gnome org>
> Cc: GNOME Desktop <desktop-devel-list gnome org>,
>      Federico Mena Quintero <federico ximian com>, foundation-list gnome org,
>      Alan Cox <alan lxorguk ukuu org uk>
> Subject: Re: Certification for GNOME apps
>
> Yesterday at 21:54, Andrew Sobala wrote:
>
> > On Wed, 2005-07-13 at 20:42 +0100, Alan Cox wrote:
> >> On Mer, 2005-07-13 at 16:27, Federico Mena Quintero wrote:
> >> > Level 2 - the app is actually written with GTK+.
> >>
> >> Why does this matter ? Surely it is about degrees of integration and HIG
> >> compliance.
> >
> > I agree. I was similarly surprised by (on the wiki) the requirement to
> > use .glade files as a possibility for one level; surely this
> > certification should be about the user experience, not coding practises.
>
> It indirectly affects many things.  Gtk+ and Glade using applications
> have a better chance of having consistent user interface AND
> translations.  Maybe it would be Gnome-certified on a lower level, but
> if it's not using stock menu items, and I have no power over managing
> it's translation, I wouldn't certify it as "fully Gnome" since it
> wouldn't fit on the desktop otherwise.
>
> Of course, there are counter examples such as Adobe Acrobat Reader 7.0
> which use Gtk+, yet don't make use of any stock labels and icons if I
> remember correctly.
>
> Therefore, using Gtk+ is surely a one level higher (think theming,
> a11y, i18n...) than not using it.  Yeah, we can make that a general
> statement such as "does this and that", but in practice, it would mean
> "uses Gtk+".  Why pretend?

I live in blissful hope of projects like metatheme and others to someday
allow a well written QT application that follows freedesktop.org standards
to properly interegrate into a Gnome desktop with users being none the
wiser.

Describing these requirements in a slightly more abstract way make it
possible for other languages and toolkits to potentially meet the
requirements (Java, QT, Mozilla XUL, who knows.)

The European Computer Driving License (ECDL) may in practice be dominated
by Microsoft but because they took the care to make their specifications
abstract we now have Linux and OpenOffice based versions of ECDL courses
being created and I'm really glad they had the foresight to make that
extra effort.

I realise it is still ridiculously optomistic on my part but I believe it
is the right way to create a certification system.

> Lets at least make certification process simple, practical and
> efficient.

I hope we can do both but if we cannot have it both ways being more
practical is important.

> > Basing the certification on library usage could lead to a situation
> > where the certification only matters to someone thinking of hacking on
> > the project!
>
> Agreed, but how to solve above issues then?  I gave one suggestion,
> but I'm not so sure it's the right one.  Shouldn't we try to be as
> practical as possible at this stage of the game?

I trust Federico can come up with good ways to balance our different
concerns.

Sincerely

Alan Horkan
http://advogato.org/person/AlanHorkan/