[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Arguments for basing upon Debian

Hi. Not sure if this is the best list to post to, but it seems the most
suitable I've seen. Please advise if you can think of any more suitable
and I'll repost there. I'm not subscribed to this list, so please CC me
to any replies :)

First, some background:

I've become very fond of Nokia's Debian-based platform known as "Maemo"
(http://en.wikipedia.org/wiki/Maemo). It's had 6 major releases to date
since 2005 and My N900 running Maemo 5 is so inspirational to me as a
developer that I love it to bits (aside from the minority of closed
source components)...but that's mostly beside the point.

For those not aware, Nokia decided to collaborate with Intel in 2010 to
try and spread the load (heaven forbid they perhaps open sourced their
mostly-close UI layer and let the community contribute *sigh*), and this
was planned by merging the Debian-based Maemo with Intel's Fedora-based
Moblin (http://en.wikipedia.org/wiki/Moblin), the result of which was
MeeGo (http://en.wikipedia.org/wiki/MeeGo). MeeGo is much more
Moblin-based than Maemo based...and this made me a little bit sad. Nokia
got a new (ex-Microsoft) CEO, who almost certainly promptly killed the
MeeGo collaboration and decided on Windows Phone instead as their
platform going forward.

Due to this, Intel has recently dumped MeeGo itself and partnered with
Samsung et al. to create the Tizen (http://en.wikipedia.org/wiki/Tizen)
project instead, which by the sounds of it is going to be very much a
clone of HP's WebOS (HTML/JS-based apps, etc).

Now, the point:

There are a number of potential projects being set up to create a
community-managed replacement project without the corporate ownership
issues that have plagued these former projects and resulted in their
constant rebranding and loss of direction.

Now, I'm a big Debian fan anyway, but Debian's extended project
stability and architecture for handling multiple platforms seems a
perfect fit for me as a base platform to base any distribution on, let
alone a mobile-orientated distribution on.

However, when making this case I seem to be constantly met with a
barrage of statements that Debian is unsuitable. I've had dismissive
statements thrown at me that .RPMs are better than .DEBs, that yum is
better than apt, and how the repository structure is unsuitable for
"mobile platforms" where there will apparently be "thousands of new apps
each day". I even read one individual stating "Linux Foundation should
be stronger and bring different players together, and shouldn't accept
excuses like for example what Debian&Ubuntu gives against them starting
to use rpm and retire deb". The "eloquence" of that individual aside,
it's very difficult to have an reasoned discussion it seems.

The end result being that most seem keen on continuing to base
themselves on OpenSuse/Redhat. I suspect this all stems down from
hostility to Nokia's treatment of the Meego platform and thus a desire
to sever all ties to Maemo...but I really do see it as a crying shame at
wasting the resources and experience that the Debian project has to offer.

So basically, what I'm looking for is more evidence that I can use in
these discussions to help make my case, as apparently Debian's resources
and experience aren't apparently enough due to unspecified technical
limitations of the support infrastructure that make it unsuitable for
mobile devices. All I can discuss personally is that I've never had a
broken system using Debian - I switched from Redhat 8 to the pre-release
version of Debian Sarge in 2004 for precisely that reason. I'm also
extremely fond of the Debian packaging process, but really don't have
enough experience with packaging for Redhat to make a convincing
argument one way or another. Personally I believe them to be near as as
good as one another (has any attempts been made to reconcile them by the
way? - Merging the best of both would seem to be a good way forward?),
but it's the project backing I think will be far more valuable.


Reply to: