[maemo-developers] Updating the info for Extras-devel non-free

[From: Jeremiah Foster]

On 11/25/09 4:07 PM, Andrew Flegg wrote:
> On Wed, Nov 25, 2009 at 14:55, David Greaves<david at dgreaves.com>  wrote:
>    
>> Does the community really have so much spare resource that we can QA
>> non-free (and presumably non-community) apps?
>>      
> fms/RST38h's emulators are non-free. However much I'd prefer to have
> the source and not special case them, they are useful packages and the
> author's intention should be respected.
>    
I don't think the author's rights trump the community's rights.
> Do I want him to go off and create a new repo? No.
This is not necessarily what will happen if we ask for source.
>   Is Ovi an
> alternative? We don't know yet.
>    
That seems to be a one way street - they can take maemo.org packages, 
but we have no access to their repos.
> However, having tested earlier versions of VGBA, iNES and others, I
> think I can say they went through the normal testing procedure,
> despite being non-free.
>
> The reason pre "-testing" they went directly to Extras was that there
> was no point going through the auto-builder. Now, however, I think
> they should be going directly into "-devel" and promoted into
> "-testing" and Extras proper as per free packages.
>    
Maemian is designed to peer into debs, to find 'bugs' in packages, not 
in software, it is designed to be part of the QA process after the build 
on garage. It will not run on debs that haven't gone through the build 
system. So at least part of the QA process will not work on non-free apps.
> Whether they get highlighted in a different queue is an interesting
> question; but will probably push non-Ovi, non-free apps away into
> their own repositories.
One of the things that the maemo version of debian's popularity contest 
is hopefully going to do is to check the repos being used on the device. 
This might show us how widespread the use of outlying repos is. It might 
also be an effective way of piercing this argument, I don't think it is 
as widespread a practice as it used to be.
>   The point of community QA is to make sure only
> good apps get to users: we're doing it because we're selfish. It's not
> free bug finding for commercial software teams; and so saying "we're
> only go to QA it for you if you give us the source" would seem to be a
> change in the purpose and intent of the QA process.
>    
I think you make a powerful argument that the QA process is for users to 
get good software. However, software developers target distributions 
like debian because of its quality assurance, and if Maemo gets a 
reputation for quality, then one can expect that sort of targeting to 
occur there as well.

It is perfectly reasonable to expect that those who participate in the 
QA process adhere to the same principles of openness that have made 
maemo.org so successful.

Jeremiah