Excerpts from Paul Wise's message of maj 1, 2018 9:10 am:
I have realised I haven't found the motivation & time to work on Debian's derivatives efforts in recent months and have been less active in recent years so now might be a good time to discuss the future of these efforts.I prefer this thread to be solely from Debian's perspective. I'll start another thread on debian-derivatives for the other side. To that end, some questions: Are folks aware of our derivatives efforts?
Yes! It has been a tremendous help in my work encouraging derivatives to [reduce/eliminate] their delta from Debian.
First concrete use of mine was in 2011, during a 3-month journey in [Asia] sponsored partly by Debian (thanks!). I met with developers of the India government-sponsored BOSS Linux, but unfortunately we had only one day together, so it was extremely helpful to be able to go through http://deriv.debian.net/BOSSlinux/patches/ with them, to point out concrete examples of things they could consider do differently, either to more efficiently share their innovative work upstream with Debian and others, or simply to reduce maintenance for them.
Since then I have passionately recommended derivatives to register with the Derivatives Census, and explore the computed patch set, whereever I could - in particular during my later journeys in [Peru] and [India], and most recently in my current employment at Purism SPC.
[reduce/eliminate]: https://wiki.debian.org/DebianPureBlends[Asia]: https://web.archive.org/web/20130101160016/http://wiki.jones.dk/DebianAsia2011
[Peru]: https://couchdesign.dk/peru [India]: https://couchdesign.dk/india
Do we need formal efforts around derivatives at all or should we just let things happen (or not) organically?
Yes, we need this. It is exactly the structured approach that I have found so useful in my work: Each project models their work slightly differently, and Derivatives Census helps in comparing across these differently modelled moving targets.
I find esp. the current opt-in approach of the Derivatives Census good. Derivatives typically want full autonomy - and only later, maybe, wants to realign with Debian.
Has Debian found any value in our derivatives efforts?
Yes. Not only have I found value personally in my Debian work (see above), but I choose to believe that Debian has also more concretely enjoyed upstreaming efforts thanks to the Derivatives Census work which would have otherwise not been done.
Which aspects of them would Debian miss if they went away?
Debian would - I choose to believe - miss patches, now passed upstream by those derivatives who wants to avoid drifting too far from Debian but not having the disciplin to track their deviations on their own.
I would - as Debian developer - miss a practical tool to compare Debian with its derivatives and help them improve in their passing their innovations upstream (see my journeys above).
More concretely, the parts I have appreciated mostly so far are the package lists and patches e.g. http://deriv.debian.net/BOSSlinux/diff_source_packages and http://deriv.debian.net/BOSSlinux/patches/ - and to some extend the wiki pages https://wiki.debian.org/Derivatives/Census/* themselves.
I am sure there are other great uses that I myself simply haven't realized yet (i.e. I wouldn't exactly _miss_ them - so arguably off topif of your question ;-) ).
Are there any changes we should make to our derivatives efforts?
Gather geolocation for derivatives with a geographic scope, and geo-mapping (i.e. visualization e.g. on a slippy map) - both to encourage collaboration across derivatives, and to aid "missionaries" like myself organize journeys.
Are there any things we should prioritise for our derivatives efforts?
Aspects that helps the derivative see themselves in comparison with Debian and other derivatives.
I am aware that this thread is about the Debian point of view, and I agree that we should discuss details of the derivative POV separately. My point mentioning it here is that Debian benefits from derivatives telling about themselves, and we best make them tell about themselves when the setting we provide them to do so is to their own benefit: I have had difficulties convincing derivatives to do the effort of "reporting to Debian what they are all about" - especially the ones that are succesful (which arguably is the ones we want info from the most).
So I look forward to the companion thread at the census list :-)
Are there any volunteers to join the derivatives efforts?
I volunteer to setup and maintain a slippy map, if that is of interest. Already have similar setups I can adapt, so should not be too much work.
I looked briefly at the scripts to maintain patches, and can maybe offer to help there - if Perl and shell code is ok (I notice both being present but Python being dominant, but didn't see any notice whether there's a preference or any code is equally appreciated for that).
Are there any volunteers to organise derivatives events at DebConf18? I'd love to have some feedback on the new derivatives page on the website and if we should have links to that anywhere on the site? https://www.debian.org/derivatives/ Any other comments?
Thanks a lot for all the work establishing and maintaining the Derivatives Census!
When first looking at it (in early 2011, I think) it felt too formal to me - intuitively it felt unwelcoming to me. I have since, through my hundreds of meetings with derivatives, learned how a more casual approach can be easily misunderstood and effectively *less* welcoming. Amazing work!
- Jonas -- * Jonas Smedegaard - idealist & Internet-arkitekt * Tlf.: +45 40843136 Website: http://dr.jones.dk/ [x] quote me freely [ ] ask before reusing [ ] keep private
Attachment:
pgpfgKrnc70EQ.pgp
Description: PGP signature