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

[Michael Niedermayer]

On Sun, Apr 21, 2024 at 05:11:36PM +0800, Rémi Denis-Courmont wrote:
> 
> 
> Le 17 avril 2024 21:58:32 GMT+08:00, Michael Niedermayer <michael at niedermayer.cc> a écrit :
> >Hi all
> >
> >The pace of inovation in FFmpeg has been slowing down.
> >Most work is concentarted nowadays on code refactoring, and adding
> >support for new codecs and formats.
> 
> OSS projects of age similar to FFmpeg are either mature (like FFmpeg), or more or less dead. Besides, FFmpeg is an established brand, which means that it's expected to be good at what it's been doing... and not expected to be good at anything else.
> 
> Of course there are also specific aspects: back then, every company made up its own codecs. Nowadays, there's at most three tracks (for video): H.26x, Chinese AVSx and AV-x, while AVC or HEVC have become "good enough" for most applications.
> 
> If (generic) you want to work on radical innovation, I think you will be better served by creating a new project. Both the FFmpeg project structure and brand would probably do you a disservice otherwise.

I will disagree on this a bit

If we for a moment look at the commerical world (but its not fundamentally different in OSS)

Projects/Companies are created to fill some need, initially they
often need to concentarte on a narrow market because they dont have the
resources to do "everything" and if they try they go bankrupt.
Once they are established and have the resources they grow or they die

Microsoft started with a OS in 1985, added an office suite in 1990
internet explorer in 1995, xbox in 2001, Microsoft Azure in 2008
and you can fill in more.
Today Microsoft is one of the largest companies in teh world.

You can do the same with apple, google, or others.

OTOH pick any company of your choice that did not expand and compare.
for example kodak and not expanding out of analoge photogrpahy is an example

FFmpeg has over a billion users indirectly. Dont you (plural) see the
opertunity here to leverage this ?

Sure this examples are commerical companies and we are OSS.
But really its the same. A company lives and dies with its
revenue and profits in $.
OSS lives and dies with its users and developers.

developers largely dont find maintaining a "mature" codebase interresting
they do find it interresting to develop new things.

And what you wrote above
"FFmpeg is an established brand, which means that it's expected to be good at what it's been doing... and not expected to be good at anything else."

yes. Iam not suggesting to have ffmpeg the applications suddenly do something entirely different
the same way as office doesnt suddenly become a internet browser.

What iam suggesting is that we should expend beyond ffmpeg, ffplay, ffprobe
beyond what our libraraies provide. And to leverage our quite large userbase

i know people dont like it but its a good example. If we added a ffchat
it immedeatly would have a large number of users.
It matters for a chat app, that there are people you can chat with havng the
same app.

You can pick something else, ffserver2 is a good example.

ffedit, a video editor for example

The situation is just bizare.
People complain on one hand that we lack new blood, we lack developers to maintain
the codebase. But then things that would bring in new blood and developers are
immedeatly opposed by 3 times more developers than are actually actively
working on FFmpeg.

We can and probably should switch to something more flashy than a pure ML
based development model with no good GUI access but thats not going to
make maintaining mature software sexy for new developers IMHO

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Rewriting code that is poorly written but fully understood is good.
Rewriting code that one doesnt understand is a sign that one is less smart
than the original author, trying to rewrite it will not make it better.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20240421/5776ccf3/attachment.sig>