FFmpeg's Leader Resigns, Hopes To Make Libav Developers Come Back

[sdack]

So first you say both projects need to be dropped, then you say ffmpeg devs should go to libav. Does not compute.

You're for some reason desperately trying to argue that libav is the supposed champ. But such contradictions in one and the same post will not convince anyone. In fact, your posts indicate that the pride and selfishness is on the libav side, completely contrary to your original claim.

You also have not addressed in the slightest the point about libav being objectively the lesser codebase and how doing it your way would require fixing bugs that were already fixed once and re-implementing decoders and filters that were already implemented.

No. First I was talking about team work and using diplomacy, which is why I called them group X and Y so that you get an understanding of what team work is about. Then I was referring to the projects specifically and the history they have. The developers already did work together and it has failed. You do need to account for it. It then has nothing to with championship. It is about making steps towards each other and to avoid for history to repeat itself.

Again, if it is unthinkable for ffmpeg developers to join the libav folks, then you will never see a combined project again. You might as well try to combine the political left with the political right, or catholics with protestants, etc..

And why is it so important that these two join? You do not see people crying about GCC and llvm not joining, or Windows and Linux. Instead are we cool about it, because it means we have choices and are not being dictated into accepting a single choice. We accept that people are different, that they have feelings, and that they are not a bunch of corporate drones. It is what has made open source strong.

[Gusar]

It is about making steps towards each other

[...]

if it is unthinkable for ffmpeg developers to join the libav folks, then you will never see a combined project again.

So it is about "making steps toward each other" but then supposedly the only possibility of reunion is that ffmpeg devs join libav, not the other way around. Again, does not compute.

You claim "diplomacy", you claim "feelings" and such, but at the same time you have a hard stance of "ffmpeg devs need to join libav, or bust". That you can't budge from this stance means all the diplomacy stuff in your posts is just meaningless fluff.

Also, you still did not address the technical stuff about the two codebases.

[yash]

If Debian switched from libav to FFmpeg, how is it a fail for FFmpeg

[Scimmia]

Again, if it is unthinkable for ffmpeg developers to join the libav folks, then you will never see a combined project again. You might as well try to combine the political left with the political right, or catholics with protestants, etc..

And again, you need to take a long, hard look at the future of libav. Right now, I see none. That's why it's unthinkable for ffmpeg devs to join it, who would join a doomed project?

Edit: Let's be realistic here. Libav has survived this long because of Debian. Debian didn't switch for any technical reason, but because their ffmpeg maintainer was one of the dev involved in the split. Without Debian, libav is nothing.

[wagaf]

And why is it so important that these two join? You do not see people crying about GCC and llvm not joining, or Windows and Linux. Instead are we cool about it, because it means we have choices and are not being dictated into accepting a single choice. We accept that people are different, that they have feelings, and that they are not a bunch of corporate drones. It is what has made open source strong.

You sound like a fool here..

LLVM is not a GCC fork. Windows is not a Linux fork. Even Google rebase their Android patches over the vanilla kernel regularly because it would be too expensive to maintain their own fork. FFmpeg and LibAV are the *same project* with the same goal, the same license, the same codebase. Having both is a complete waste of time and energy for many people every day...

[sdack]

And again, you need to take a long, hard look at the future of libav..

No. It only ends being the same horse shit, which has been coming from Michael Niedermayer, when it is for the developers of each project to decide what they want to for their future.

If you want to know my opinion on what would be the best outcome then I can tell you... The best would be for both groups to keep cracking on, not to be influenced by a resigning leader (after all is he sort of taking the blame for it), for ffmpeg to find a new leader, and then to start thinking about the possibility of making a new project. Why? Because most open source projects evolve over time anyway. Newer projects appear, older ones disappear and are being superseded by newer ones. So instead of insisting on a Cinderella story with a Hollywood happy ending, lots of tears and hugging among developers, would a new project with a new goal be the smarter solution. It should not be about fixing the past, but about making the future.

To elaborate on this further, for example OSS was replaced by ALSA, which has been replaced by PulseAudio, should both groups combine their efforts in a next generation project. People are already making an effort towards a so called "PulseVideo", which I personally dislike. Better would be here a new platform like a "PulseAudioVideo" (or whatever you want to call it), because audio and video belong together more often than not. The technical challenges are then found in achieving a perfect synchronicity between audio and video and so the logical step here is to process both in a single process, created by a single project, and not in separate processes coming from separate projects. I see the future not in PulseAudio or PulseVideo, ffmpeg or libav, but in a project, which combines them. A software, which allows users to play and record audio and video, to use many formats for a maximum in compatibility, usable on any hardware from mobile phones to PCs, with a wide support among distributions and support for as many A/V devices as possible. It should be about enabling users to create audio and video and to let them share it with one another, whenever and wherever. The easier this gets the more the users will benefit from it. This is my opinion as a user and not as a developer.

Now I am sure someone will tell me that this is their point when they demand for libav to merge back into ffmpeg, but I will keep telling you that it ignores the fact that this is how it started, and that clearly it failed. There is then no point in doing the same thing again and hoping that the second time around it is not going to happen again. Instead, the personal relationships between the developers need to be accounted for just as one needs to account for changes in a video format, or as for changes in the compiler and what it is the users need. The only way one can hope to create a new dream team is to give everyone a fresh start with new goals, a new project and a new leadership. It will still allow for people to bring in old grudges, but then at least they all have new things to focus on. Just trying to do the same thing again when it has failed means that one did not learn from it.

[sdack]

You sound like a fool here..

LLVM is not a GCC fork. Windows is not a Linux fork. Even Google rebase their Android patches over the vanilla kernel regularly because it would be too expensive to maintain their own fork. FFmpeg and LibAV are the *same project* with the same goal, the same license, the same codebase. Having both is a complete waste of time and energy for many people every day...

It is not about labels. LLVM is not a fork of GCC, and still there is plenty of talk about how LLVM is the better project and how GCC developers better be working on LLVM. Such talk is just as wrong. What right do you believe you have to demand a reunion? Because it is a fork? There is no such clause! There is only the right to create forks, to modify the code and to use it in a new project.

Also it is not your time that is being invested here. You have no say over who is allowed to spend time on which project. The libav people have already decided a long time ago that working with the ffmpeg project is a waste of their time, which is why they took the initiative to create a fork. If anything then their decision deserves respect.

Is it not bad enough when millions of people hate their jobs and bosses, then still drive to work every day, because it pays the bills and feeds their families? To expect the same obedient attitude from the developers, many of whom are not being paid for their work and at the same time create brilliant pieces of software, is unrealistic. It is because of the freedom that comes with open source that people participate without payment and that makes it the better software.

Whatever argument you think you may have you better think again. Open source projects are influenced by those who contribute to them, but little by those who complain about them.

[geearf]

If you want to know my opinion on what would be the best outcome then I can tell you... The best would be for both groups to keep cracking on, not to be influenced by a resigning leader (after all is he sort of taking the blame for it), for ffmpeg to find a new leader, and then to start thinking about the possibility of making a new project. Why? Because most open source projects evolve over time anyway. Newer projects appear, older ones disappear and are being superseded by newer ones. So instead of insisting on a Cinderella story with a Hollywood happy ending, lots of tears and hugging among developers, would a new project with a new goal be the smarter solution. It should not be about fixing the past, but about making the future.

Hmmmm.... why do we need something newer than ffmpeg (or libav)? What is ffmpeg unable to do that a newer project could?

To elaborate on this further, for example OSS was replaced by ALSA, which has been replaced by PulseAudio, should both groups combine their efforts in a next generation project.

Hmmmm again...
PA did not replace ALSA, it sits on top of it, without ALSA there is no PA.
But neither ALSA nor PA forked from OSS so I am not sure how that relates anyway.

[toyotabedzrock]

New slogan for LibAV, We never let security come before our personality conflicts.

[sdack]

Hmmmm.... why do we need something newer than ffmpeg (or libav)? What is ffmpeg unable to do that a newer project could?


Hmmmm again...
PA did not replace ALSA, it sits on top of it, without ALSA there is no PA.
But neither ALSA nor PA forked from OSS so I am not sure how that relates anyway.

We need something newer, because we always want something newer. And yes, PulseAudio did replace ALSA from a user perspective, which is what I wrote.

I know that it can be hard to accept as a geek, when you have finally mastered the art of ffmpeg after months of learning, when you know every of its options by heart and have become proud of yourself for all the things you know about it, but this is not the future of it. The future does not lie in command-line tools. It will be tools with minimal, but intelligent interfaces, which provide auto-detection of formats and devices, that anticipate the user's goal and guide them through it by predicting parameters. Tools that make it possible for anyone to do what a 25-year old geek can do now.