Which distro


I'd also have to say I think we should go with Debian. I've been running Ubuntu for a few years now, long enough to see features I like disappear, each upgrade with Canonical calling it a UI improvement. It's completely put me off installing any new versions of Ubuntu.

- Neelix

Edit: @Perusa That video contains no useful information whatsoever.
 
Last edited by a moderator:
Just to make things clear,  using ubuntu does NOT mean using any of the canonical crap.

Look at xubuntu/kubuntu/mint , they're all based on ubuntu, but they don't like the canonical stuff any more than you do.

They're just using it as a more up to date debian to base their distro on.

If we go with debian, testing is probably the best option.

That has more mature arm support than stable, and more recent packages.

Thing with debian testing however, is that you can't be sure on how it will behave, what they'll throw in/out, and when.

Although I don't know how much of an issue this will be, I never ran testing for any extended period myself.

Using an existing distro doesn't give you the amount of control you had over things as something you completely built yourself, like we did with angstrom.

At least with the upcoming LTS version of ubuntu, you'll know there won't be any unexpected changes for the next 5 years.

(In case updates do cause more problems than expected)

However, this still doesn't mean I'm completely pro ubuntu now, in fact the test image I just built was based on debian unstable :)
 
For the Testing or Stable question I got another idea:

When you first start your pandora, it asks you (among others), what Desktop Environment you want. You could also ask, whether the user want stable or testing with a little tiercet of explanation.

Both have Pros and Cons, but I think for most users stable is fine.
 
At least with the upcoming LTS version of ubuntu, you'll know there won't be any unexpected changes for the next 5 years.

(In case updates do cause more problems than expected)
Debian Jessie will be frozen at the end of this year. Perfect at the earliest time we can count on the Pyra!Even if a new Debian version is released ~2 years later, Jessie will still get updates, which means no unexpected changes for at least (!) the next 4 years, too, starting with 2015.

So in fact it has at least the same period time as with Ubuntu 14.04 (because the support period starts in 2 months, which is quite useless for the pyra, if it gets released at least in one year...)
 
Last edited by a moderator:
The idea is to have one standard install + toolchain that all .pyra (.pnd replacement) will work on.

If any "advanced" users want to run something else, that's fine, but "officially" supporting multiple versions of firmware isn't really an option.

Especially not with binary distributed apps.

(Where "multiple versions" are completely different versions of distros, not different package installed on top of a "pyra-base" selection from the same repos)

Ubuntu and debian do have nearly the same method of building packages and images, so if all the pyra support apps are written cleanly (and they should be), making the actual choice can be delayed a bit untill the device itself is ready for release.

Im mostly focusing on debian now, because that's what I have running here anyway, but we should look at any advantages an alternative might have.
 
I'd also have to say I think we should go with Debian. I've been running Ubuntu for a few years now, long enough to see features I like disappear, each upgrade with Canonical calling it a UI improvement. It's completely put me off installing any new versions of Ubuntu.
What he says. Exactly.

Although I could be swayed to some sort of Arch thingamadoodle.
 
Don't forget the psychological component. I don't think an "open" console sells very good to linux users if the default operating system is ubuntu.

And linux beginners or Windows users you can tell, that the default OS is very much like ubuntu, that it is in fact the basis of Ubuntu, but with special changes for the Pyra. ;)
 
Hmm, something I don't quite know: If we have .pnd-next files with either .deb dependencies or completely self-contained ones.

If they are built to work on Debian unstable, those with .deb dependencies would mostly not work on Debian stable, would they? The self contained ones would?

If they are built to work on Debian stable, those with .deb dependencies would work on unstable? The self contained ones too?

Are there answers to those questions which can generally assumed to be true? Even in some years when there is a new unstable/stable?

If that wouldn't cause problems with .pnd-next files, I would suggest this: Ship with Debian stable. Anyone who wants more recent packages can easily upgrade to testint/unstable.

From what I read, unstable seems better than testing to me: http://www.debian.org/doc/manuals/debian-faq/ch-choosing.en.html
 
Ship with SuperZaxxon and PND system just like the Pandora now.  Whoever wants to run something else they could do that on there own.  Like all my computers come with Windows preinstalled but I got rid of Windows once they got to my house.  It's not too hard to replace an OS on a computer these days.   Problem solved.
 
From what I read, unstable seems better than testing to me: http://www.debian.or...hoosing.en.html


3.1.5 Could you tell me whether to install testing or unstable?

This is a rather subjective issue. There is no perfect answer but only a "wise guess" could be made while deciding between unstable and testing. My personal order of preference is Stable, Unstable and Testing. The issue is like this:

  • Stable is rock solid. It does not break.
  • Testing breaks less often than Unstable. But when it breaks, it takes a long time for things to get rectified. Sometimes this could be days and it could be months at times.
  • Unstable changes a lot, and it can break at any point. However, fixes get rectified in many occasions in a couple of days and it always has the latest releases of software packaged for Debian.

I nearly forgot about that bit for testing.

And we really can't afford any breakage at all for a built in firmware.

So it's either :

debian stable, with older packages, and support for arm that isn't quite there yet.

or

ubuntu based , with more up to date packages and better arm support.

A third option would be to base it on debian testing or unstable, but run our own mirrors of everything and only update our repos if there's nothing broken.

(which would mean a lot of extra maintanance work, not to mention a huge amount of space, so I don't really see that going anywhere)

Also,we really want to give everyone as good an experience as possible, not just give you a device with a useless firmware that won't run anything at all, and .pyra packages that need to include loads of extra libs to run.

Things should "just work".  This isn't for linux experts only.
 
Last edited by a moderator:
Debian stable or if we are close to a next stable release, debian testing that will become stable a few months later.

For exemple if the Pyra is available at a time when the drivers work better on testing (it is quite common with recent hardware in Debian), we could decide on the testing version, but as a Debian version, not as eternal testing (using the Toy Story name of the release in the sources.list). So when that release becomes stable, the OS automatically becomes the stable release.
 
debian stable, with older packages, and support for arm that isn't quite there yet.


The raspberry pi uses Debian as it's official OS. I have had debian running on my Dockstar for 3 years without any issue. Those are ARM devices, what exactly is lacking in Debian on ARM ?
 
I was disappointed when Pandora went with Ångström and from the beginning I thought they should have used Debian or Ubuntu.  I remember it was said that neither had near the support for ARM that Ångström did but in the long run it would have been better going Debian or Ubuntu.  I don't know though, maybe using Ångström was better, who is to say.

This time around lets go with either Debian or Ubuntu.  I think in number of desktop users Ubuntu is the number 1 distro and Debian is number two so for the community size they can not be beaten, by anyone.

Also, going Ångström didn't really get us many Linux devs where if we went Debian might have gotten a huge number of Debian ARM developers buying the Pandora just to tinker on.  Ubuntu has more users but not really any more developers and I think we'd get more developers with Debian as some Debian devs won't touch Ubuntu because of their non-existent upstream patches.

So Debian or Ubuntu, I think more Debian.
 
Back
Top