Re: SFO call for action - aka questionnaire

[Mike Turquette <mturquette gmail com>]

On Mon, Jun 29, 2009 at 8:48 AM, Nils
Faerber<nils faerber kernelconcepts de> wrote:
> About you:
> 1. Please describe your background a little bit, i.e. what are you
> working on that made you interested in Gnome/mobile? If it helps us
> understand your background, who are you working for?

Hello!  I'm a Linux Engineer at Texas Instruments, working on our
OMAP3 and OMAP4 chips.  I've personally been a GNOME user on the
desktop for many years now, but my real involvement GNOME *Mobile*
started when I wanted to replace the Busybox userspace used on our
OMAP development platforms with something more modern, something
closer to what our customers were shipping (ie, Nokia's Maemo).  I
replaced our internal build system with Poky Linux and as such have
been diving into the world of OpenEmbedded in parallel with exploring
GNOME Mobile.  I want both to succeed and be completely awesome.  The
responses in this email are as personal developer and not
representative of TI.

> When you first heard of Gnome/mobile:
> 2. What did you think it to be? Please be as frank as possible - what
> were your really *first* thoughts?

I already knew it would be a collection of parts to be assembled, just
like desktop GNOME, and I knew it would not be a Linux distro like
Poky, Moblin or Maemo.  I did expect there to be a bit more formality
about 3 things:

1) specific listing of what those parts are
    the closest thing to that is the old block diagram on
gnome.org/mobile or perhaps the mobile directory on the GNOME release
ftp server
2) how to assemble those parts in a GNOME Mobile fashion
    IE, is matchbox wm assumed, is it assumed that there is no OpenGL
and only OpenGLES, etc?
3) and related to #2, what constitutes a distro/device being GNOME
Mobile conformant

> 3. What did you expect from it? What did you hope to see as a result
> from it?

I hoped rapid prototyping of devices and distros would be quicker.  I
wanted to start combining geoclue with EVERYTHING!  Also I hoped a
strict list of desktop libs that were considered non-Mobile would be
maintained to prevent things from getting too bloated.  I was
basically expecting a rulebook for how to keep a distro/device GNOME
Mobile-compliant, while allowing development to happen really quickly.

This was not the case and OE turned out to the fastest way to
prototype, and Poky let me do it in a GNOME Mobile-ish fashion
targeted at small touchscreen devices.  GNOME was always my first
choice, but if Poky had not existed and some KDE/Qt solution did then
I would have gone with that as second choice.  Business is business
you know ;-)

I also hoped the project would help traditional desktop-only GNOME
developers to keep mobile in mind while writing apps/libs.  I still
believe this is vital for success: we're not going to succeed by
running some bizarre fork of desktop GNOME, but instead by educating
the vast majority of desktop developers and utilizing a subset of
their work.

> 4. Why did you join the group? Do you feel like a "member" at all? If
> not, why not?

I joined the group because I personally buy into the idea of GNOME
Mobile.  However I don't feel like a member.  I feel like a guy
watching other members who don't talk to each other while voicing my
opinion in an attempt to get said members to talk.  I'm familiar with
the GNOME community and process, so hopping on this mailing list and
spitting out a few opinions wasn't that hard for me.

> 5. Did it match your expectations? If not, why? If yes, how?

Its been more quiet then I had hoped.  This stems from a lack of
direction and a clear idea of what GM actually is.

I've found the most success by supporting Stormy and her Netbook
Usability Study.  TI donated two Zoom2 platforms for testing (and
later for development by as-yet-undecided GNOME Mobile developers).
Shameless plug:
http://www.logicpd.com/products/development-kits/texas-instruments-zoom%E2%84%A2-omap34x-ii-mdp

> 6. Is there something like a top-three of most important goals that you
> (personally, not thinking of the group as a whole) would like to see
> addressed by Gnome/mobile? (Three would be great, one is enough too)

1) component-level (aka project-level) description of what makes a
valid GM distro/device
2) some program to educate desktop GNOME devs on how to write apps and
libs that aren't totally useless to mobile
3) a GNOME Mobile czar who actually has a real idea of what the
project should be and can talk with device manufacturers, software
shops & the community to grow this project.  A mailing list and a wiki
page isn't what made the larger GNOME project what it is today...

> And today:
> 7. How do you perceive Gnome/mobile today? What is missing? What is too
> much? What is / could be Gnome/mobile's first priority goal? And second?

Realistically I see GM as an idea and a possibility more than a real
product.  When I look at the existing GM distros out there (lets use
Poky, Maemo and Moblin as examples) I see two basic themes:

The first type of distro (Poky) would be doing just fine even if GNOME
Mobile didn't exist as a project.  The infrastructure/libs used are
mostly just desktop ones, but just a really limited subset are used,
with an eye towards memory consumption.

The second type of distro (Moblin/Maemo) also don't need the GNOME
Mobile project to exist either, because they basically wrote a ton of
their own infrastructure code (I'm looking at you, Hildon, Clutter and
oFono!).  Like the first type of distro the existing GNOME
infrastructure used was basically just desktop stuff, but used in
small amount (no libbonoboui, no libgnomeui, etc).

In both cases the applications on top of this infrastructure are
almost all entirely new and written per/device, so there is once again
no reuse of anything the GNOME Mobile project provides.

> Technically:
> 8. What do you think should qualify a platform for being a Gnome/mobile
> platform? A specific set of components used? If yes, which components?

I agree that GM-conformance should be defined by components.  I
haven't written a list yet, but maybe I will!  This question requires
more thought than an off-hand response.

I do feel that 3D stuff might need to become a requirement though,
which probably is controversial.  Requiring OpenGLES and
Metacity/Clutter would be a fascinating direction for the project to
go.

> 9. Or should there be a formal qualification process? If yes, what would
> be your idea about it? If not, why not?

Best to have a project with real deliverables before we go creating a
conformance test and a committee to govern the committee tasked with
making that conformance test ;-)

> 10. What should be the technology goals of Gnome/mobile? Like creating a
> toolbox rather than a product? Or collaboration and information
> infrastructure rather than a development project?

I think we should follow desktop GNOME's lead with this: I think that
desktop GNOME is mostly a toolbox + basic components, but using
jhbuild it is easy to make a product.  Likewise GM should focus on
tools + components, but perhaps using jhbuild or OE it should be
relatively simple to create a vanilla GM product without any of the
differentiation that device manufacturers are going to put in their
real products.

Whether this results in code getting written for the GM project or
perhaps guidelines mandated to traditionally desktop-focused
developers is a question of what the members of this project WANT to
do.  Its all pretty much volunteer after all.

> Organisation:
> 11. At the moment the Gnome/mobile group is a pretty loose group,
> everyone interested is welcome and can at once participate. Would you
> like to see this a little more formally structured? Like a formal
> membership, application process, levels of membership, voting processes,
> etc.? Or would like to see it grow as is, a group connected by ideas and
> consensus rather than rules? Other ideas?

A truly personal question!  I'm a "show me the code" kind of guy, so I
don't care about any formal process for application and membership.  I
hate to see good work/ideas excluded because they come from a
non-member, likewise I hate to see bad code/stupid ideas accepted just
because they came from membership.  I say keep things informal for
now.  Lets build software first, organizational bloat second.

> Comments:
> 11. Anything else? Did we forget something? Any additional comments?
> Thoughts?

Thanks for putting this together Nils!  Hopefully my answers are
helpful (and not too negative)!  I really like GNOME Mobile and want
to help it succeed.  Linux may never dominate market share on the
desktop, but the mobile space is ours for the taking and GM has a real
shot at being used in many real world products.

Regards,
Mike Turquette

> Again, many many thanks in replying and for your time and work to craft
> your answers!
>
> Cheers
>  nils faerber
>
> --
> kernel concepts GbR        Tel: +49-271-771091-12
> Sieghuetter Hauptweg 48    Fax: +49-271-771091-19
> D-57072 Siegen             Mob: +49-176-21024535
> http://www.kernelconcepts.de
> _______________________________________________
> mobile-devel-list mailing list
> mobile-devel-list gnome org
> http://mail.gnome.org/mailman/listinfo/mobile-devel-list
>