[FFmpeg-devel] [RFC] 5 year plan & Inovation

[Andrew Sayers]

On Sat, May 04, 2024 at 09:28:03PM +0200, Michael Niedermayer wrote:
> Hi
> 
> On Fri, May 03, 2024 at 03:45:20PM +0000, Cosmin Stejerean via ffmpeg-devel wrote:
> [...]
> > What doesn't exist (yet) is a way to keep people on the exact email based workflow 
> > we currently have, and have bi-directional sync with something like github or gitlab.
> > Such a thing could probably be built, but it might be worth first trying to see if those
> > that insist on sticking with the CLI can use one of the existing CLI based workflows.
> 
> Such a thing could be quite useful to many more projects than just ffmpeg.
> There are many older projects that use ML based workflows.
> 
> I imagine STF might be willing to fund such a thing if it is technically
> feasable. As the goal of STF is about maintainance. And bridging the gap
> between old ML and new browser based workflows allowing developers who
> prefer to work through their web browser to do so.
> 
> also, we need to find maintaince related projects worth minimum 150k €
> for 2025 for STF.
> We cant do many of the things we do in 2024 for STF again as they where
> one time things and STF doesnt like sponsoring adding new features.
> 
> thx

It seems like the strongest argument for sticking with the ML is from
experienced maintainers who don't want to jeopardise their existing
workflow; while the strongest argument for switching is from people
itching to try out new workflows.  So how about this for a plan...

Make a repo on SourceHut, not necessarily for FFmpeg itself but for
automated review tools (running fate tests, checking C11 compliance
etc.).  Their CI/CD system automatically runs those tests on every
patch, then we manually forward genuine issues to the ML.  That would
let experimenters show off new things, and would let maintainers
think through what their workflow would look like in a mixed
environment.  Then when we've got enough evidence to make a long-term
plan, we can wind the repo down without too much fuss.