[Fso] FSO Revitalizing Plan

Denis 'GNUtoo' Carikli GNUtoo at no-log.org
Fri Mar 14 20:13:40 CET 2014


On Thu, 13 Mar 2014 12:29:46 +0100
"Dr. Michael Lauer" <mickey at vanille.de> wrote:

> Hi folks,
Hi,

> So my plan to revitalize would be like that:
> 
> 1.) Copy the relevant repositories to github.com as a synchronized
> clone of git.freesmartphone.org
Well, some people (me included) don't like github that much, because no
source code is provided for their service.

> 2.) Make sure that what we have there
> builds against recent tools (autofoo, vala, 3rd-party dependencies)
> 3.) Apply outstanding patches. 
Good, If I understood well, JaMa kept the whole stuff compiling, and
there are still some users reporting the status in the StagingTest
page[1].
So while they test on SHR, it also applies to FSO.

> 4.) Wade through our spamfull trac,
> copy those tickets that are still relevant to github and shut down
> trac completely. 
Well, the trac could be made read-only with a big warning saying that
the new instance is for instance at redmine.freesmartphone.org

> 5.) Ponder about the current state of FSO on recent
> and future target devices, think about extensions and enhancements.
For devices, I don't think it's worth to do any new port but for
devices like the GTA04, and the neo900.
In SHR/FSO it always went like that:
1) some people/users/developers start buying the same device
2) the device is new, they realise it's way faster then the previous
generation and get very exited about it.
3) they work hard on it at the beginning.
4) and step by step, they work less hard on it.
5) Then the port is nearly finish, like remains very few things to do
to get it usable for telephony and basic use case(some people would
have been very happy with it).
6) At that point all developers already switched to a new device and
went back to 1)
7) Some people still try to fix all previous devices that were
abandoned just before succedding, and try to do the "very small thing
left", but many things prevents them to do it fast:
 a) there are many devices that are into the same situation, so you
 multiply the work by the number of devices.
 b) some developers are not available, and they're the expert on the
 areas that need to be fixed.
 c) many bugs also need to be fixed, like for instance vala+efl broke
 again.

So when we look at what's left now, remains only the GTA02, and the
GTA04. So maybe we should think at why theses two devices still remain
while all the other were abandoned.
The GTA02 is the most supported phone, becuase at the beginning there
were a lot of developers working on it, the downside is that it has a
very old kernel, and I wonder how systemd can still work on it.
Like the GTA02, the GTA04 also got lots of developers at the beginning,
and its kenrel is recent now.
FSO didn't start at the beginning of the GTA02, so the vala
implementation got started very late.
At that point the GTA02 was really well documented already.
And No reverse engineering or anything close to it was needed at that
point.
The GTA04 got less developers, but still a big number compared to other
phones(like the nexus S, htc dream etc...), and it didn't require any
reverse engineering or anything like that.
It only needed some adaptation of FSO and of the kernel, for issues
that were not covered at all by the current frameworks(in the kernel
and in fso).
Beside the GTA02 and the GTA04 are GNU/Linux phones. Maybe it's related
to that too:
Let's take someone who has a Nexus S, he buys it with Android.
Then if he installs SHR, it will not work well, because the power was
not finished yet.
If he buys it for SHR, he has to be a developer willing to work on it.

For the pure software part (like what stack is desirable on a GNU/Linux
phone), we should find what's good and what's bad about FSO, and look
at similar libraries and compare FSO with them.
We could for instance reuse some of them as an FSO backend, or replace
parts of cornucopia that weren't written yet(opimd).

We should also keep in mind the number of developers that will be
involved and how soon they want something that works.

> What do you think?
Personally, I always had a dylema while participating to SHR:
The freerunner is the best phone I know freedom-wise.
Then comes the GTA04, and maybe next the neo900[2].
FSO runs on SHR and on debian.
SHR was way more usable than debian: it contained the lastest
enlightenment, which was way more finger friendly, and the lastest FSO
with all bugfixes.
If I fixed a bug, I had it fixed in the official images too very soon.

Now the dylema is related to the fact that neither SHR, nor debian were
considered 100% free software.

When I used SHR, many bugs needed to be fixed to remain usable, given
the number of bugs, I had no time left to make it 100% free software.

And I had a hard time using debian on it.

That is to say, I whish that FSO succeed to be revived, and maybe one
day a 100% free distro could use it and be usable on some phones(like
GTA02, GTA04, neo900 etc...).

References:
-----------
[1]http://shr-project.org/trac/wiki/StagingTests
[2]I'm unsure if all the components have been decided but there may be
some issues with the camera firmware, or the lack of separate GPS(So
the user would have the choice between not using the GPS(thanks to an
antenna switch) or giving the modem access to the GPS.

Denis.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: not available
URL: <http://lists.goldelico.com/pipermail/fso/attachments/20140314/25fe397c/attachment.bin>


More information about the Fso mailing list