Well, i intended to write a quite comprehensive answer to this question with an detailed elaboration of determining factors, my motivation and the "historical background" of my deceision process, but somehow i have not managed to do it within the past months...
But now that i am about to release TAK 2.0 i had to do something. So here is an excerpt of my post in the Newer version of FLAC? (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=76733&view=findpost&p=672649) thread, that gives at least a short answer:
(jcoalson @ Dec 10 2009, 23:39) *
at this point, given thomas' stated goal to try and benefit somehow from all his hard work, my advice now would be to keep it closed and stay in the lead in its niche. once it's open, any practical advantage it has over flac could be added to flac with less effort than it would take existing devices and software to switch to tak. to open it up, you have to be completely resigned to not getting anything for it, because that's what may happen.
...
Besides this i again have to agree... I only want to clarify (for users not familar with TAK's development history) that when i was talking about possible benefits i meant non-material ones.
Unfortunately the validity of thoses statements creates kind of a deadlock situation for TAK: If it wants to attract considerably more users, it has to go open source, but at the same time it does, it's advanced technology can be copied and it will be only a matter of time, until TAK is dead...
Among other important things, this is the primary reason why i haven't worked on an open source release yet. Don't think this was always clear to me, it was merely a notion i somehow avoided to realize, because it would have killed much of my motivation.
Honestly, why should i put effort into an open source release, if this only dispatches TAK!?
No. This would be self-denial.
Currently the only valid option for me to reveal TAK's technology would be:
- Improve the codec until i am running out of new ideas.
- Then ask Josh, if he would like to put the TAK codec into FLAC and if there will be a proper reference to me.
It is your software, your hard work and your choice.
I think your work will be remembered if it is in TAK or part of the next generation of FLAC. Hopefully Josh would give you the credit you deserve if your work is incorporated into FLAC.
If TAK is open sourced, maybe TAK could be the "experimental" codec, free to grow and change, without fear of breaking compatibility, while being a live test bed for possible FLAC features, pushing the older codec to be better.
First, whatever you do with the code source in the future I want to thks you for finally trying to solve the mysterie of TAK's licence.
Personnaly, I don't know what advice to give to you, because as a possible future TAK user, I am undoubtly biased. It is an evidence that any end-user would benefit from having the code source open, but in the same time it is an evidence that going open source is a loss of control for any developper.
The actual situation was not sane. Personnaly your hesitations reminded me of X-Chat (http://xchat.org/) which was an open source IRC client which suddenly became a shareware on win32 while still being open source on linux. The situation is different with TAK but for the end-user, it is the same, the mixing of license creates a no-man's land where no end-user want to step. When X-Chat became a semi-open-source software, I switched to chatzilla & I never used X-Chat again. I had the same feelings when I tested TAK without knowing its license.
I understand that going open source is not an easy task, as there are a lot of FUD (http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt) surrounding the question. Here on HA, old users will remind of what happens Mr Questionman: if I recall correctly within a week after Gambit released the source code another guy added a few features & claimed the software his own, which disgusted Gambit of open source for a long time.
This kind of shit happens everyday, I cannot tell you there is no risk:
1: you cannot go back if you change your mind (X-Chat).
2: you can be stolen (Mr Questionman).
For point N°1, you have to decide for yourself & be 100% sure you won't change your mind, I cannot help you, all I know is that having a schizophrenic developer is bad for software adoption.
For point N°2, well all we can do is to make the robbery public & put the thief to shame, ... honestly I don't even recall the name of Mr. Questionman hijacker ... but I do miss Gambit here on HA. For me it is not the same to be stolen by a moron who wants a minute of fame & to to "stolen" by another open source software due to the march of technologie.
I recall that Josh asked you to work together in the beginning of Yalac (I don't recall if the offer was formal or in between lines) because he was foreseeing the actual fate of TAK, you cannot complaint now that all your work will be lost when you declined the offer. It is not a shame to the Aoyumi of FLAC. You claimed you had the shoulder to do it all by your side ... & now that you're so close you seem to hesitate as you realize that end-user may not follow you ... or at last not instantly. Once again over the years you're not logic. But anyway it's better to realize late that Josh had some insight than never.
The real question is where do you want to go from now. I think TAK as reached the point where everyone know that you are a skilled coder & that if you endlessly improve TAK & swallow all Flac features (including a friendly licence) ... at some point you may take the crown.
The difference between now & the yalac days is that you realized that even with a ultra-optimized-codec the switch between flac & tak will not be easy & will take years.
So the question is: are you ready to assume the support for TAK for as long as TAK is not perceived as a standard ? Once again I cannot answer for you, but all I know is that if the answer is "no" then TAK was a real waste.
In fact you don't really a have hundred of choices:
- You release the source code & support TAK for your lifetime & maybe TAK will become a standard.
This solution requires a lot of work on your side but make everybody happy (... except Josh maybe).
- You release the source code & stop supporting TAK, well this is a pity as you could have given the TAK tips & tricks to Josh a few years ago. It is unsatisfaying for you, but all is not lost for end-user & you're still a major contributor to a major software, even if you're not the boss anymore.
- You don't release the source code & TAK is doomed to be a benchmarking codec for a niche of enthousiasts. This is the worst solution as you will leave a bitter taste in the mouth of users who trusted you in the yalac days.
No one can decide for you, but plz once you decided hang on to your choice, because your hesitation is what makes the difference between David, Josh & you.
If you decide to release the source & support TAK, it means you take the lead & people will ask you to act as a leader which means supporting the weight of responsability.
First, whatever you do with the code source in the future I want to thks you for finally trying to solve the mysterie of TAK's licence.
I couldn't put you in jeopardy to get warned again... Earnestly, like i wrote earlier, you were right to request an answer.
Sorry, if i don't yet comment on any of your statements, i am just a bit tired.
I recall that Josh asked you to work together in the beginning of Yalac (I don't recall if the offer was formal or in between lines) because he was foreseeing the actual fate of TAK, you cannot complaint now that all your work will be lost when you declined the offer. It is not a shame to the Aoyumi of FLAC. You claimed you had the shoulder to do it all by your side ... & now that you're so close you seem to hesitate as you realize that end-user may not follow you ... or at last not instantly. Once again over the years you're not logic. But anyway it's better to realize late that Josh had some insight than never.
Some corrections:
- I offered Yalac to Josh.
Addendum: And we were into email contact. At some point i recall some misunderstanding (maybe on my side), not by any means important, but in combination with other things as lack of time on my side, we lost contact for a while. And then i became aware, that i woudn't have enough freedom to improve the codec (in incompatible ways) as soon as it had been adopted by FLAC. So i didn't contact Josh again.
- I am not complaining about something like a "lost chance". If TAK/Yalac had been adopted by FLAC when it was in it's early state, many improvements would never have been implemented. Not to forget: I would have had less fun. For me TAK development was (and will be for some more time) adventure and exploration with some unexpected turns. You can't do it this way if you want to establish a standard as Josh did.
If i have regrets about something, then about my inexperience when i arrived at hydrogen. I was overhelmed by the positive feedback and wanted to fulfill any user requests. (Seems as if i am now beginning to write the missing parts of the comprehensive statement i actually wanted to make...) "Open source, no problem!" But i didn't think about the consequences. Forgive me, but this was also my first participation on any forum and i didn't know how it could affect me.
But i was very soon aware of TAK's position. I knew that it could only fill a niche like WavPack or Monkey's Audio, what is not really bad if you can make some thousands of users happy.
I never thought it could acquire FLAC's position, at least not without a tremendous amount of work. And here we are talking about a kind of work i don't want to do, for instance marketing.
If you decide to release the source & support TAK, it means you take the lead & people will ask you to act as a leader which means supporting the weight of responsability.
I woudn't like to be a leader unless there isn't a better choice (someone who likes this position)... That's not the point. I like to be an inventor.
Thomas
edit: See
Addendum above.
Sorry for bad English, English is not my native language.
I'm not only writing for Thomas, but to everyone who're asking for Open Source.
Please respect the author, don't push anyone to release the source code, when they feel the time was right, they'll release the source, I've seen someone ask the author in a very rude way, it's like a robber trying to ask for your wallet.
And ... personally I feel codec like TAK have no need to Open Source NOW.
Codec like TAK usually complex, and have a lot of special tuning by the author, few people will invest the time to understand or improve it (well ... good buys exist but only a few), I'm not sure how many people help Josh to improve FLAC, but I'm sure not too many ... a project is much easier to handle by a single person than a lot of people, and to cooperate developers need extra time and effort.
Appreciated to those help to test and feedback ideas, but I've seen too many people just plain talk, without donating a single dollar, just shouting they need Open Source.
Maybe time can be spent in a better way, such as coding front-end or utilities for the codec.
Sorry if the language was too strong, I mean no offense.
- I offered Yalac to Josh.
Oh ! My memory is playing tricks on me, sorry ... I am gonna kindly slap Josh then
a niche like WavPack or Monkey's Audio
I agree WavPack fills a niche market, but Monkey Audio is much more popular than just a niche, HA doesn't reflect the worldwide reality. Despite being slow & closed Monkey Audio attracts a lot of lossless beginners just because of its compression & GUI. Monkey Audio is on the decline but compared to wavpack or TAK, its user database is still huge.
WavPack has always been a great codec but back in the time when it could have been very popular wavpack 3 was missing instant seeking so in that time people favored Monkey Audio. Then when wavpack became instantly seekable it was too late Flac was rising.
IMHO you are over-estimating how flac is a de-facto standard that would be undethronable. Sure Flac is forever rock stable & perfectly usable, but don't trust the HA zealots blindly... a few years from here the same kind of zealots were spreading the musepack supremacy to the world. They were wrong.
If wavpack & monkey audio didn't succeed to become standards, it might simply due to the fact that (with the exception of max compression achievable) both of these codec are inferior, simply. With his exotic features like hybrid or self extractable .exe David never really cared about wavpack being the "so-called" best anyway. I hope wavpack, flac & tak, can all 3 co-exist as open source codecs & let the end users decide what is best for them.
There is a paradox IMHO within the claim that you just want to be an "inventor" & the fact that you fear to release the code ... let's look at David as an exemple instead of focusing on Josh: the exotic features of wavpack makes it unique (even if IMHO not optimal for backup due to its slowness). His hybrid mode makes him a real "inventor". Now look at the license he has chosen: BSD. For me David is rationnal to himself as he has chosen from the start that wavpack was non-profit free time hobby project & he has chosen the license in consequence. You claim that you only want to be an inventor, but in the same time you prevent the world from fully enjoying your invention. For me you are not as logic as David. For me it is obvious that you want something more than just being an "inventor", but I don't really know what because we have nothing to offer for your code except our sincere thanks.
There is a paradox IMHO within the claim that you just want to be an "inventor" & the fact that you fear to release the code ... let's look at David as an exemple instead of focusing on Josh: the exotic features of wavpack makes it unique (even if IMHO not optimal for backup due to its slowness). His hybrid mode makes him a real "inventor". Now look at the license he has chosen: BSD. For me David is rationnal to himself as he has chosen from the start that wavpack was non-profit free time hobby project & he has chosen the license in consequence. You claim that you only want to be an inventor, but in the same time you prevent the world from fully enjoying your invention.
Don't speak on behalf of the world. Thank you.
Releasing source code is a one-way street as someone has mentioned. One cannot un-release source code.
Almost nobody cares if TBeck happens to be less decisive or less rational than David, because both of them have created codecs that their respective niches fully enjoy.
If TBeck makes using TAK very restrictive or ceases to support it, people will simply jump to their next best choices Wavpack or FLAC. These people have a unique set of requirements which TAK currently fulfils best overall.
If you don't like the fact that TAK isn't going to be open source "
now-damnit-I-say-now!", it apparently does not fulfil your set of requirements, and you are free to pick from other equally excellent lossless codecs that are free for use (both free as in free beer and free as in not free beer). TBeck has explained his decision and substantiated it with his reservations. That was what you wanted from him; the explanation, and I certainly hope you're not demanding anything more, especially since it is most likely outside your rights to do so.
You claim that you only want to be an inventor, but in the same time you prevent the world from fully enjoying your invention.
Nonsense. I'm enjoying Tom's invention right now (yes, I'm not the world, but also many people in the world don't have computers either).
I don't buy this "open-source or die" deal.
Anyone can write a bespoke license that allows and disallows whatever the author wants. If Tom offers a very good SDK and allows manufacturers to decode TAK files then doesn't this open up platforms and opportunities? Surely this is possible without open sourcing? I may be wrong - I'm not a licensing expert, but I don't see why a win-win solution, whereby Tom maintains his rights and the codec spreads is impossible (especially where fees are not the issue).
Does open-sourcing a decoder ultimately give away the encoder too? I presume it does.
C.
I already said several time that the problem with TAK was not that it wasn't open source, but that it didn't had a clear license.
Plz stop accusing me of being an open source inquisitor, I don't use linux & I never will & Richard Stallman (http://en.wikipedia.org/wiki/Care_Bears) is not my friend.
If Tom wants TAK to be closed source, it is his absolute right. But then it is my absolute right to use FLAC. Many people seems to read this statement as an utimatum ... this is not.
This is just my truth about the situation of TAK for me. I don't think I am alone in this situation, even if indeed I cannot speak on behalf of the world.
Even if TAK stays closed source, its speed/efficiency improvement alone might be enough to justify the switch for my personnal use. So accusing me of being a free software zealot is a non-sense. I do favor free software when I have the choice, but my HDD is full of patented & closed technology ... it doesn't prevent me to sleep. But even closed source I want Tom to clearly tell the world what is the TAK license for the next ten years.
Plz stop accusing me of being an open source inquisitor
No one did. The only thing I took issue with you personally was your peculiar statement I quoted. The fact you decided to take the rest of my post (which did not speak to you directly, and was mainly a series of questions) personally is equally peculiar.
C.
The thread is called "When will TAK go open source?", but it doesn't even answer the question "Whether or not will TAK go open source". If we were dealing with a corporation, i would assume the PR department is doing a great job.
Yes, going open source is a one way street. As far as i am concerned, keeping codec secret is a valid (but poor) choice. But keeping a future license secret for such a long time is a one way street also. Personally, i made a decision - i won't use or support a project, which is managed in such a way.
Despite being slow & closed Monkey Audio attracts a lot of lossless beginners just because of its compression & GUI.
IIRW the Monkey Audio source was opened long time ago but I haven't hear of somebody's contribution so far. I might be wrong, though.
It's not just a binary decision...
You can have an open "standard" without going open source on your implementation.
You can have a patent, and still open the source.
You can do something with the encoder, and something else with the decoder.
Rather than worrying about someone taking over your project, you can take over someone else's (e.g. branch FLAC).
There are lots of options open to you.
In the end, though I've both created it and used it, the world of open source software is generally quite sad (in the traditional sense of the word): lots of people doing great work, but not getting paid for it. Unless you have a day job or sponsor, and "spare time", it's not something you can keep doing.
Oh, and either bevery careful with your choice of licence and terms, or use this one...
http://sam.zoy.org/wtfpl/ (http://sam.zoy.org/wtfpl/)
...because I've seen too many nasty arguments about open source software licences!
Cheers,
David.
а.п.т.:
I am no authority on the matter but as far as I understood some part (maybe all, I dunno) of Monkey Audio code was unofficially "open" (by "open" read "available on demand") but Monkey Audio remained closed via its license. As far as I understund 3rd party developpers didn't like this situation much either. I recall I heard complaints there too, but it was clear that Monkey Audio was not going open source & that it was only a "trick" to improve Monkey Audio support. <== Warning, some part of the above might be wrong as I never used Monkey Audio myself.
Edit: I think I recall there was a 3rd party command line Monkey Audio encoder at some time, I dunno if it still exist, but I think I recall the complaints where coming from there ... I admit it's still very shrouded in my memory.
2Bdecided:
http://sam.zoy.org/wtfpl/ (http://sam.zoy.org/wtfpl/)
That f#cking license is f#cking freedom, I f#cking love it
Does open-sourcing a decoder ultimately give away the encoder too? I presume it does.
Nope. RAR is an example. It has open source decoder with licence clearly stating that it can't be used to reverse the encoder.
I already said several time that the problem with TAK was not that it wasn't open source, but that it didn't had a clear license.
Well that was the case with me. When I first saw TAK, I decided to wait until it's sources are opened.
@TBeck:
The answer by 2Bdecided is the one you should put more attention in.
When one develops a codec, not only develops a software, but also a format. Both parts (the code and the format definition) can contain IP (Intelectual Property).
You have to think what use you expect your program to have. For example, if you expect it to be used on mobile devices, you will need to code it yourself, or ask someone to do so (this still doesn't require opensourcing).
You did a step in this direction creating the decoder dll, which then has been used to create plugins for some players. But this still limited it to Windows.
I don't know how much of the performance of TAK is thanks to your implementation, or thanks to the format you have defined. Publishing the format would allow others to try to do their own implementations, probably not as fast as yours. The example of vorbis and aoyumi is a good example, and recently, FLAC is getting new implementations that improve it ( for example, flacuda or fpFLAC ).
In the end, what can make TAK to persist is not your .exe, but the fileformat.
The development in the way you've managed it, gives yourself freedom to elaborate and try different things, while being in control of it as a whole. This part wouldn't be affected by open-sourcing it, but probably wouldn't benefit of it either.
I have been developing an audio program for eight years, and I wasn't its original author. I've always done it in my free time, and we had good and bad times. Being opensource is not the major factor, but thanks to being open source, I could be the maintainer of it, as well as the rest of the people that has come and added their part into it.
TBeck:
Just one question ... if you offered Yalac to Josh, why do you fear that releasing TAK's source code would kill TAK because its technology would be swallowed/copied by Josh ? There is something that I am missing there Maybe there is a rational explantion, but said like this it sounds weird
The thread is called "When will TAK go open source?", but it doesn't even answer the question "Whether or not will TAK go open source". If we were dealing with a corporation, i would assume the PR department is doing a great job.
While i understand that i deserve criticism for waiting too long with an definite answer (and deceision), i don't know what is unclear about this:
Honestly, why should i put effort into an open source release, if this only dispatches TAK!?
No. This would be self-denial.
Currently the only valid option for me to reveal TAK's technology would be:
- Improve the codec until i am running out of new ideas.
- Then ask Josh, if he would like to put the TAK codec into FLAC and if there will be a proper reference to me.
What else can this mean as "No, TAK will not go open source"?
And oh, i called the thread "When will TAK go open source?" Maybe because this was the question people asked in the past and therefore this seemed to be the most appropriate title?
TBeck:
Just one question ... if you offered Yalac to Josh, why do you fear that releasing TAK's source code would kill TAK because its technology would be swallowed/copied by Josh ? There is something that I am missing there Maybe there is a rational explantion, but said like this it sounds weird
If you are talking about a possible future corporation: Why should i put effort into a source code release of TAK if it's only purpose would be to make it's technology available to FLAC? It would make more sense to modify the existing code to be easily adoptable by FLAC. And then only the core of the codec would be required, not the much bigger source for reading, writing, streaming, command line management and so on.
edit: Maybe i haven't understood your question right?
- I offered Yalac to Josh.
Oh ! My memory is playing tricks on me, sorry ... I am gonna kindly slap Josh then
No need to.
I have just filled another gap in my second post:
Addendum: And we were into email contact. At some point i recall some misunderstanding (maybe on my side), not by any means important, but in combination with other things as lack of time on my side, we lost contact for a while. And then i became aware, that i woudn't have enough freedom to improve the codec (in incompatible ways) as soon as it had been adopted by FLAC. So i didn't contact Josh again.
Maybe i haven't understood your question right?
Well ... no ... but maybe that's because I didn't fully understood what you what you meant by " I offered YALAC to Josh" in the first place. Edit: Your "Addendum" mostly solved it.
Anyway your last answer to Gregory S. Chudov answered most of my incomprehension toward TAK license.
I wish you the best with your codec. Personnaly I will keep using flac & encode some of my files to lossy (likely nero aac) in order to save some space.
Again thks for the time spent trying to explain where you were going with TAK, it was usefull as now I know that TAK is not for me. It might sound sad, but it is not. It is more of a relief sensation, because the future of my HDD is clearer. I guess I personnaly don't like experimental codec much because within a month I discarded both lossywav & TAK in a row for my own use.
I still wish that one day I can use some of your code in FLAC, as, from my point of view, becoming the Aoyumi of FLAC is the best thing that can happen to you & your code.
If it ever happens it will be the dream codec that I was expecting TAK to become. Honestly I don't hold my breath as my experience tells me that it will never happen, but I cannot blame you as I don't know what I would do myself if I were in your position.
Anyway your last answer to Gregory S. Chudov answered most of my incomprehension toward TAK license.
Huh, so my initial statement was less clear than i thought... Sorry!
Does this apply to the decoder too? Having TAK in ffmpeg eventually would be very handy, and would allow for portable/embedded support.
Good point. If decoding source code was available, TAK playback support could be implemented in linux, and encoding could be done with wine. Not ideal, but still better than nothing.
Good point. If decoding source code was available, TAK playback support could be implemented in linux, and encoding could be done with wine. Not ideal, but still better than nothing.
Far from ideal. I do think you have hit the nail on the head though. Surely the most important next step for TAK is not whether the source should be opened (and only the dev can decide that), but making native binaries freely available for the end-user on multi platforms.
On a technical front, TAK is the most advanced and impressive lossless audio codec imho. You've only got to start using it and the results speak highly. But I don't use it because I use Windows, Linux and OS X.
If the Windows binaries (encoder, decoder, plugins) were available natively on all 3 platforms, it's a no brainer, I would switch. Would I care if it was open source? Not really.
I think you've hit the nail on the head there. The license of the source code is not the issue at stake here; the availability of encoder/decoder binaries on alternate operating systems is.
I am still not sure why should TAK go Open Source. There is absolutely no reason for this. Does it work? Yes it does. The only issue I see is linux/mac version, which is, as I see it, nonexistent.
But there are other codecs available in that world, so I don't see it as such a big deal. So, where exactly is the problem?
But there are other codecs available in that world, so I don't see it as such a big deal. So, where exactly is the problem?
No problem. The world will keep spinning, indeed. Would you say the same if TAK was Mac-only though?
It's not just about Linux and Mac OS X versions, it's about support on any software and hardware platform that Thomas does't provide. If TAK were Free Software, or at least open source with reasonable licensing, third-parties would have the opportunity to add such support themselves. It doesn't always happen (Ogg Vorbis' acceleration patches have long stayed Windows-only), but what matters is the opportunity. I can't bitch about FlaCUDA being restricted to Windows right now, I can only hope Gregory (or yet another third-party) will port it over to my platform of choice (linux). Given the availability of CUDA under linux and FLAC's popularity, I wouldn't be surprised if someone jumped on the occasion once they learn about FlaCUDA's mere existence.
By keeping TAK closed source, Thomas keeps the door closed to a whole bunch of talented and motivated programmers who would gladly port his codec and/or improve on it. Guys like Gregory Chudov (FlaCUDA), Justin Ruggles (Flake), GeorgeFP (fpFLAC), Aoyumi (Vorbis), Jason Jordan (shnutils), others I didn't catch the name of (Vorbis, Lancer) and many others I can't think of right now.
Edit: and not least, Josh Coalson (FLAC). I'd mention other authors like David Bryant (WavPack), though I don't know if they'd benefit from TAK, since I don't recall statements to that effect from them. Their case is also a bit different since they're the authors of their own codec.
Unfortunately the validity of thoses statements creates kind of a deadlock situation for TAK: If it wants to attract considerably more users, it has to go open source, but at the same time it does, it's advanced technology can be copied and it will be only a matter of time, until TAK is dead...
You are a pessimist. You still can open the decoder only for now (maybe without a no-reconstruction-of-the-compressor-restriction known from UNRAR ???), you can publish more detailed specs, and you also can contact Xiph if they would adopt it as they did with FLAC - more people would be available to defend if against possible pirates. TAK has been around here for 4 years now, so hardly anyone can steal it and say "it's my work"
No problem. The world will keep spinning, indeed. Would you say the same if TAK was Mac-only though?
Well, yeah. I tried it a few times, and continued to work with FLAC and APE. And a lot of software is Mac-only, and I find the alternatives. As well as for some linux software.
It's not just about Linux and Mac OS X versions, it's about support on any software and hardware platform that Thomas does't provide. If TAK were Free Software, or at least open source with reasonable licensing, third-parties would have the opportunity to add such support themselves. It doesn't always happen (Ogg Vorbis' acceleration patches have long stayed Windows-only), but what matters is the opportunity. I can't bitch about FlaCUDA being restricted to Windows right now, I can only hope Gregory (or yet another third-party) will port it over to my platform of choice (linux). Given the availability of CUDA under linux and FLAC's popularity, I wouldn't be surprised if someone jumped on the occasion once they learn about FlaCUDA's mere existence.
I don't see this "free software" as something that should be applied to all software. I use linux sometimes at home, and much more at work, I used it for a long time as desktop OS - and I understand the willingness of linux users to have everything for free - and for the most of the time, the software is free. But!
Thomas decides to keep the TAK format for himself. So what? You can do two things in this case, first one is to ask him
politely to make linux and os x binaries, if he can and have knowledge of doing that, or continue to use some other lossless format. Why attacking him, and forcing him to do something he doesn't want to do? I am sorry you can't use TAK except from within wine on linux, but that is
your choice, not his. Do you really think he should do something he doesn't like because you choose to use OS he doesn't support? Linux is all about choice, so you have the choice of some other software to use.
By keeping TAK closed source, Thomas keeps the door closed to a whole bunch of talented and motivated programmers who would gladly port his codec and/or improve on it. Guys like Gregory Chudov (FlaCUDA), Justin Ruggles (Flake), GeorgeFP (fpFLAC), Aoyumi (Vorbis), Jason Jordan (shnutils), others I didn't catch the name of (Vorbis, Lancer) and many others I can't think of right now.
So? I see him as a talented coder/programmer. His choice is to continue to work on codec for himself. You are linux user, you see what happened to kernel just because everyone can contribute all that is seen by me, as a bystander, is attack on his decision not to release the codec source code. Please, respect his wishes. Just because people on this board helped him test the codec it doesn't means we have the right to choose what will become with that codec. It's as simple as that.
And I am sorry if that offends someone, but that is the way I see things.
Tom, there seem to be a great many positive and pragmatic contributions in this thread:
1. OPEN SOURCE DECODER ONLYDoes open-sourcing a decoder ultimately give away the encoder too? I presume it does.
Nope. RAR is an example. It has open source decoder with licence clearly stating that it can't be used to reverse the encoder.
You still can open the decoder only for now (maybe without a no-reconstruction-of-the-compressor-restriction known from UNRAR ???), you can publish more detailed specs, and you also can contact Xiph if they would adopt it as they did with FLAC ....
Does this apply to the decoder too? Having TAK in ffmpeg eventually would be very handy, and would allow for portable/embedded support.
2. MULTI-PLATFORM BINARIESIf the Windows binaries (encoder, decoder, plugins) were available natively on all 3 platforms, it's a no brainer, I would switch. Would I care if it was open source? Not really.
3. ALTERNATIVE APPROACHES
You can have an open "standard" without going open source on your implementation.
You can have a patent, and still open the source.
You can do something with the encoder, and something else with the decoder.
Rather than worrying about someone taking over your project, you can take over someone else's (e.g. branch FLAC).
There are lots of options open to you.
----------------
Tom, I have 2 questions regarding TAK:
1) Is there a problem with (or good reason for not) open sourcing (only) the decoder?
2) Hypothetically if the TAK decoder was open source, and a hardware (e.g. DVD Player) manufacturer was deciding to impliment FLAC and/or TAK compatibility, are there any additional obstacles that would make TAK less attractive to them than FLAC, specifically in regard to the licence and the code?
C.
I don't see this "free software" as something that should be applied to all software. […] I understand the willingness of linux users to have everything for free […]
You sound confused. Free Software is capitalized for a reason. It doesn't stand for free as in beer, but free as in speech. Free to modify and distribute. Americans sometimes borrow the french
libre to avoid such confusion.
ask him politely
[citation needed] Was I ever impolite with him?
to make linux and os x binaries, if he can and have knowledge of doing that
It's not that he can't or refuses to do that; by his own admission, he's overworked as it is. That's what's most frustrating: he won't give someone else the chance to take on that task in parallel to his own work.
Why attacking him, and forcing him to do something he doesn't want to do?
Who's attacked him? Who's forcing him to do anything? Certainly not me. That's a groundless accusation.
I am sorry you can't use TAK except from within wine on linux, but that is your choice, not his.
Sure. The difference is, the platform I use is indeed about choice, while Thomas' position is about restricting everyone to use his software as he sees fit. Yes, it's his right, just don't ask me to agree with it.
Do you really think he should do something he doesn't like because you choose to use OS he doesn't support?
I think he doesn't want to open the source for the wrong reasons. At this point we'll have to agree to disagree though.
You are linux user, you see what happened to kernel just because everyone can contribute
I don't know what you're talking about, actually. The linux kernel keeps better and better with every release - improvements that I can actually see and measure.
all that is seen by me, as a bystander, is attack on his decision not to release the codec source code. Please, respect his wishes.
I doubt Thomas himself sees it as an attack. Also, I wouldn't be expressing myself on the topic - again - if he hadn't put it on the table - again.
Several free software programs (= cost price zero) but not being open source have/had their success, for example Exact Audio Copy and foobar2000.
Everybody has to make for him/herself clear the REAL GOAL of doing something (thousands of reasons are possible). This can be a tough excersize as different feelings mixed with external factors are hard to prioritize. Once one takes a decision he/she knows the road to follow, and, there will be no change until a new decision is taken. Important: not taking a decision is in essence also a decision.
I can fully understand Thomas and can only hope that mental battles won't steal hours from improving his lifework. Thanks for sharing your thoughts.
Several free software programs (= cost price zero) but not being open source have/had their success, for example Exact Audio Copy and foobar2000.
There is a difference between closed source programs (fb2k, iTunes, WMP) and closed formats (ALAC, WMA, TAK).
Honestly, why should i put effort into an open source release, if this only dispatches TAK!?
No. This would be self-denial.
Currently the only valid option for me to reveal TAK's technology would be:
- Improve the codec until i am running out of new ideas.
- Then ask Josh, if he would like to put the TAK codec into FLAC and if there will be a proper reference to me.
What else can this mean as "No, TAK will not go open source"?
I think there is a 3rd option - a hybrid really.
FLAC, as you stated, cannot adapt and change as much as TAK. FLAC could not use all of the technology of TAK. TAK can live on, FLAC would incorporate some features, and TAK would be free to grow and evolve as you get new ideas.
You get the credit you deserve. TAK lives on. FLAC improves. Seems like the best for everyone!?
First: Surely i am following this thread, but there may be some delay before i reply, because i am busy preparing the V2.0 release.
I hope, the following will answer most new questions:
1) Couldn't you at least release the source code of the decoder?
No. Unfortunately the source of the decoder would already reveal most of the codec's technology. Most of it's compression power could then be achieved by adopting the code. Only my very fast encoding algorithms would be kept secret.
2) Does TAK's speed depend a lot on your specific implementation?
No. It's mostly based upon algorithms. For instance even a plain pascal implementation would encode -p1, which easily beats FLAC -8 compressionwise, several times faster than FLAC -8. And this although FLAC -8 is using assembler MMX optimizations.
3) Wouldn't going open source help to improve TAK's efficiency a lot?
No. Sorry, but if there are so many motivated developers waiting to improve lossless codecs, then why hasn't the very attractive FLAC project improved more efficiencywise? I can see many opportunities to improve FLAC's speed and some to improve it's compression efficiency without breaking the backwards compatibility.
Ok, there are flacuda and fpFLAC. Really nice work of the developers! But isn't it better to improve algorithms instead of applying brute force? Hey, let's save a bit of power. And how many users trust those implementations and will use them regulary to compress their possibly large file collections? I don't know...
4) Wouldn't going open source help to get some hardware implementations?
Maybe. But what's about Monkeys' audio and especially WavPack? Both are open source (ok, a bit questionable in the case of Monkey's), but are there many hardware implementations? No. Lossless audio compression generally seems to be a bit exotic, therefore the developers of hardware players wil possibly implement FLAC, but anything else is usually too exotic, unlikely to attract enough users to justify additional development costs.
5) Do you intend to release Linux implementations of TAK?
Yes. But i have to learn a lot, before i can do this.
6) Do you hate open source?
No! It's great in many cases, but not always. What i don't like are crusaders, making fantastic claims and ignoring big parts of reality or experience. I would have liked to release an open source version of TAK, but unfortunately the current setting (bad english?) doesn't make it possible. My current preference is a later fusion with FLAC.
4) Wouldn't going open source help to get some hardware implementations?
Maybe. But what's about Monkeys' audio and especially WavPack? Both are open source (ok, a bit questionable in the case of Monkey's), but are there many hardware implementations? No. Lossless audio compression generally seems to be a bit exotic, therefore the developers of hardware players wil possibly implement FLAC, but anything else is usually too exotic, unlikely to attract enough users to justify additional development costs.
This is not a good argument for 2 reasons:
1) APE has extremely high decode requirements that make it impossible to use on most embedded devices, and until very recently, there was no available embedded decoder. Its simply not analogous.
2) APE actually does have ok hardware support given the massive limitations imposed by the format. Some Cowon and Iriver players decode it, as well as rockbox. If you want to use APE with a hardware device or a stereo, you absolutely can find hardware to do it. Most other companies that are even open to FLAC simply do not ship hardware fast enough to decode APE (e.g. Sandisk).
A better example would have been wavpack, which is efficient to decode but also not widely known like TAK. It can be decoded on some Cowon players as well as Rockbox.
FWIW I think this is an odd thing to worry about. In the long run either TAK will fade into obscurity (ATRAC Lossless), or the format will be reverse engineered anyway like previous closed lossless formats (e.g. ALAC). If you have some intention to cash in or exploit your edge, you should probably consider how exactly you will do this sooner rather then later. Otherwise you may end up with the worst of both worlds (your ideas widely used by others and no one using TAK).
A better example would have been wavpack, which is efficient to decode but also not widely known like TAK. It can be decoded on some Cowon players as well as Rockbox.
Maybe that's why i wrote "But what's about Monkeys' audio and
especially WavPack? " ?
FWIW I think this is an odd thing to worry about. In the long run either TAK will fade into obscurity (ATRAC Lossless), or the format will be reverse engineered anyway like previous closed lossless formats (e.g. ALAC). If you have some intention to cash in or exploit your edge, you should probably consider how exactly you will do this sooner rather then later. Otherwise you may end up with the worst of both worlds (your ideas widely used by others and no one using TAK).
If other open source lossless codecs could only be improved by reverse engineering and stealing the ideas of others, which light would this shed on the open source philosophy?
You may as well have missed this:
My current preference is a later fusion with FLAC.
This seems to be the best option for me (if i get proper credits) and the users.
But what's about Monkeys' audio and especially WavPack? Both are open source (ok, a bit questionable in the case of Monkey's), but are there many hardware implementations? No.
I don't know about "many", but incidentally, my Cowon S9 supports Monkey's Audio in addition to FLAC.
5) Do you intend to release Linux implementations of TAK?
Yes. But i have to learn a lot, before i can do this.
Since the command-line encoder/decoder can already be used with wine, what's sorely missing is playback support. How are you going to provide that for the many music players out there, while keeping the source closed? It just won't work. You won't spend time to write a plugin for every music player, and no maintainer is ever going to accept a binary blob in their project anyway.
I wish you'd clearly state that there will be no linux support, ever. TAK would be officially a Windows-only codec, and it would put an end to all the drama.
But what's about Monkeys' audio and especially WavPack? Both are open source (ok, a bit questionable in the case of Monkey's), but are there many hardware implementations? No.
I don't know about "many", but incidentally, my Cowon S9 supports Monkey's Audio in addition to FLAC.
I know, that there are
some hardware implementations. But i doubt, that's sufficient to spread a format.
5) Do you intend to release Linux implementations of TAK?
Yes. But i have to learn a lot, before i can do this.
Since the command-line encoder/decoder can already be used with wine, what's sorely missing is playback support. How are you going to provide that for the many music players out there, while keeping the source closed? It just won't work. You won't spend time to write a plugin for every music player, and no maintainer is ever going to accept a binary blob in their project anyway.
I wish you'd clearly state that there will be no linux support, ever. TAK would be officially a Windows-only codec, and it would put an end to all the drama.
Aren't you a bit nitpicking? Would my statement only be honest if i added a list of possible player plugins and their exact release dates? Do you expect me to be a lot more precise than the vendors of paid software? Sorry, but this is still a hobby for me, and sometimes real life breaks in and changes anything.
Maybe i am a bit too open: Me saying that i deserve some criticism for my late deceision regarding an open source release seems to be an invitation for some people to slap me.
Yes, i am not a lawyer, and i am not adding a lot of pages of explainations and definitions to my posts, which possibly could make them really distinct. Hence you may always find something to complain about. Often another possibility would be to ask me.
I am wondering, how straight and never failing all those people are in their real lives, who keep on critizising me because of one (surely important) failure.
But i forgot, forums are also about fun and entertainment...
Aren't you a bit nitpicking? Would my statement only be honest if i added a list of possible player plugins and their exact release dates? Do you expect me to be a lot more precise than the vendors of paid software?
Not at all. I'm trying to make a point. I think linux support without opening up the sources is not realistic.
Aren't you a bit nitpicking? Would my statement only be honest if i added a list of possible player plugins and their exact release dates? Do you expect me to be a lot more precise than the vendors of paid software?
Not at all. I'm trying to make a point. I think linux support without opening up the sources is not realistic.
Yes, "you think". That's the point. I am making a statement and you expect me to add an elaboration of possible interpretations and implications.
If i was a politician, i would simply bring the self-responsible citizen into play.
But i am not. Therefore i can only say: My posts are mostly simply posts, not case studies.
A better example would have been wavpack, which is efficient to decode but also not widely known like TAK. It can be decoded on some Cowon players as well as Rockbox.
Maybe that's why i wrote "But what's about Monkeys' audio and especially WavPack? " ?
I did misread that particular line but it doesn't change the correctness of my argument, nor support yours, but I am sorry I poorly phrased that sentence.
If other open source lossless codecs could only be improved by reverse engineering and stealing the ideas of others, which light would this shed on the open source philosophy?
Interesting that you jumped to complaining about open source, although I did not mention it
Lots of closed formats copy the ideas of other closed formats too. The idea that somehow not allowing people to look at the source keeps people from understanding what you are doing in the long run is not particularly sound. Eventually all good ideas are copied or else obsoleted.
My current preference is a later fusion with FLAC.
This seems to be the best option for me (if i get proper credits) and the users.
I didn't miss that, I just doubt people will be particularly receptive to amending the flac bitstream, so I don't think this will ever happen. I wasn't going to be such a downer, but since you asked . . .
But what's about Monkeys' audio and especially WavPack? Both are open source (ok, a bit questionable in the case of Monkey's), but are there many hardware implementations? No.
I don't know about "many", but incidentally, my Cowon S9 supports Monkey's Audio in addition to FLAC.
a subset of monkey's audio is supported by cowon and a bunch of chinese DAPs you can't get here (they're not too worried about patents) but not much else. no hardware supports the whole format. afaik wavpack is only in the cowon players.
thomas is right that going open source is not going to automatically get you into players. wavpack has had slightly higher compression ratios and bsd licensed code for years.
3) Wouldn't going open source help to improve TAK's efficiency a lot?
No. Sorry, but if there are so many motivated developers waiting to improve lossless codecs, then why hasn't the very attractive FLAC project improved more efficiencywise? I can see many opportunities to improve FLAC's speed and some to improve it's compression efficiency without breaking the backwards compatibility.
you could still do that. I don't know what kind of credit you're looking for but all flac project work is attributed. that's why I advised the way I did in the yalac days, to contribute to flac. you get more visibility contributing to a high profile project than trying to start a competing one from scratch.
honestly I didn't start flac to get credit and I would not feel right to advance the field slightly and then close it off. I saw it more as a scientific pursuit, building on something open and keeping it open.
another misconception here is that ideas can be stolen. they cannot, they're not like bars of gold. no amount of 'imaginary property' laws can usurp natural law. you either have to keep it secret (until someone else picks the same idea out of the ether), or set it free with no attachments.
A better example would have been wavpack, which is efficient to decode but also not widely known like TAK. It can be decoded on some Cowon players as well as Rockbox.
Maybe that's why i wrote "But what's about Monkeys' audio and especially WavPack? " ?
I did misread that particular line but it doesn't change the correctness of my argument, nor support yours, but I am sorry I poorly phrased that sentence.
ked . . .
My argument was: An open source implementation on it's own isn't sufficient to spread a format and to automatically cause a significant amount of hardware implementations. That there are some implementations for Monkey's audio and WavPack doesn't contradict this, if FLAC has so many more. All are beeing open source, but only FLAC is so consequently beeing marketed, and this seems to make the difference.
If other open source lossless codecs could only be improved by reverse engineering and stealing the ideas of others, which light would this shed on the open source philosophy?
Interesting that you jumped to complaining about open source, although I did not mention it
Intersting: You wrote "or the format will be reverse engineered anyway like previous closed lossless formats (e.g. ALAC)". I only know of open source implementations of the reverse engineered ALAC, if you know better, please tell me.
And this again looks like nitpicking: Any possible facette of my statements seems to be the object of investigation of some hidden underlying motivations of mine. Please don't forget, that my english isn't very good. I could be a lot more precise in my native language.
you could still do that. I don't know what kind of credit you're looking for but all flac project work is attributed. that's why I advised the way I did in the yalac days, to contribute to flac. you get more visibility contributing to a high profile project than trying to start a competing one from scratch.
Well, i think i will contact you within the next days.
honestly I didn't start flac to get credit and I would not feel right to advance the field slightly and then close it off. I saw it more as a scientific pursuit, building on something open and keeping it open.
Me too. Otherwises i wouldn't have worked on TAK for so many years (before arriving at hydrogen) only for my own fun and without talking about it.
Unfortunately some not so nice life events have changed this to some degree. But the details -while beeing important to really understand the course of the public TAK development and my attidude change- aren't appropriate to be published.
As a software developer I know how you feel, Thomas. If I can give you some advice, discard any wishful thinking and weight the options.
If you keep TAK closed, will it really benefit you? Are you sure? If you don't think so, then you should probably make the mob happy. If you think it just might and you have a plan for how, then go ahead and keep it closed, though don't be so naive as to think people aren't gonna be angry or disappointed. We've been looking forward to TAK going open source for years and now you've gone and cancelled Christmas on us.
And the 'I just want credit' line doesn't ring true. There's no shortage of applause for open source developers. If you dream dollar bills, Tom, there's no need to hide it; we've all got some grudging respect for entrepreneurs.
My 10 cents, my 2 cents is free.
- Qest
you could still do that. I don't know what kind of credit you're looking for but all flac project work is attributed. that's why I advised the way I did in the yalac days, to contribute to flac. you get more visibility contributing to a high profile project than trying to start a competing one from scratch.
Well, i think i will contact you within the next days.
Imho is the best option Newest Flac versions (e.g.): "Flac 1.5 (TAK Powered)"
My 2 cents
An interesting read this thread is
(lol how Yoda esk does that sound!).Qest makes a good point
If you keep TAK closed, will it really benefit you? Are you sure? If you don't think so, then you should probably make the mob happy.
Sounds simple I know, but ultimately it's true. There's only so long that you'll work on TAK as real life becomes more important. And then what happens? TAK becomes a "rjamorim's ReallyRareWares" addition... which ironically no longer exists (unless there's a new url for it??).
I can't see why you wouldn't get credited for helping to improve FLAC once Josh got the code; he certainly appears to be open to that. Which is why you two should have a good chat on MSN etc and get it sorted.
You've already inferred that you never created TAK for monetary gain; so as long as you are satisfied with getting "credit" (best to be amicably and legally agreed in writing between you & Josh), then I think you'll be a happy man. You'll have more freedom in your real-life and yet as Josh also infers, you'll get the opportunity to continue improving your ideas via adding to the FLAC (or FLAC-branch) source code.
Mike Giacomelli & 2Bdecided have given some great info/advice.
The very fact you created this thread also, imo, implies you know what you want to do, but just want to air your thoughts and see what others think.
Sometimes you have to realise that things happen for a reason and often they work out for the better.
Like dB, that's my 2p
I am guessing most of your motivation comes from community recognition.
Just realise that all code you write will eventually be outdated and replaced. This can happen soon enough should your real life get in the way, as it has happened with almost every project.
What do you want to be remembered for? The author of a nice little program that was used for a short while (like Apollo, Reclock, Wingate, etc). Or as part of something bigger?
I have personally spent thousands of hours on open source projects that are quietly fading away. This is really sad for me.
You've already inferred that you never created TAK for monetary gain;
That may have been the case initially, but it is possible that he may have had a change in his income recently, as for many people these days, and he is now having to look for ways to supplement his income.
Instead of arguing about open source, lets put our heads together and come up with ways that we all can benefit, him through some income, and the rest of us through access to an excellent piece of software.
Thomas, if someone comes with a codec using an algorithm that is already in TAK can you prove that you invented it? Don't you think it would be a good thing to release the source code with a license that ensures it cannot be used elsewhere without your consent? (No I do not know which license to use )
You could ask for a small fee from each person requesting the source code. If the variables/comments are in German, so be it-- just make sure the fact is well known to anyone who requests the source code.
Thomas, if someone comes with a codec using an algorithm that is already in TAK can you prove that you invented it? Don't you think it would be a good thing to release the source code with a license that ensures it cannot be used elsewhere without your consent? (No I do not know which license to use )
The classic decision for most any IP that can remain hidden in the final product. Whether to keep it secret, or patent (which requires public disclosure). WIth "secret" you have no protection against someone either coming up with the same idea themselves or figuring it out from what's publicly observable.
If you don't think so, then you should probably make the mob happy.
Although I wasn't a follower of TAK back in the YALAC days, I have been keeping an eye on the discussions for a couple of years now. This "mob" has really only been just a couple of people; at least those who have been vocal about it in a somewhat less than respectful way.
afaik wavpack is only in the cowon players.
Out of the box maybe, but thanks to Rockbox (http://www.rockbox.org/) support can be enabled on some Archos, iRiver, SanDisk and Toshiba DAPs plus a few generations of Apple iPod as well as early Mini and Nano ones, add to that support in CorePlayer Mobile (http://coreplayer.com/content/view/28/69/), a piece of software available for the Symbian S60 3rd, Symbian UIQ v3, Windows Mobile PocketPC / ARM, Windows Mobile Smartphone / ARM, WindowsCE MIPS and Palm platforms, and finally consider the Logitech Squeezebox (http://www.logitechsqueezebox.com/) and the DViCO TViX Universal Jukeboxes (http://www.tvix.co.kr/ENG/), so still far from FLAC but not too close to Monkey's Audio either IMO.
thomas is right that going open source is not going to automatically get you into players. wavpack has had slightly higher compression ratios and bsd licensed code for years.
Right indeed, the argument is till valid, I concur.
This "mob" has really only been just a couple of people; at least those who have been vocal about it in a somewhat less than respectful way.
Thank you. I'm glad I'm not the only one of this opinion.
You've already inferred that you never created TAK for monetary gain;
That may have been the case initially, but it is possible that he may have had a change in his income recently, as for many people these days, and he is now having to look for ways to supplement his income.
Instead of arguing about open source, lets put our heads together and come up with ways that we all can benefit, him through some income, and the rest of us through access to an excellent piece of software.
the problem sounds like conflicting goals. you can get more visibility through the open source route but for a codec then you have essentially no chance to make any money off of it, ever. have to choose one, but once you do the former there's no going back.
to make money, I would say keep it closed and patent everything possible,
if you have an identified niche where you could license it. that might be what tilman liebchen did with mpeg als.
the problem sounds like conflicting goals. you can get more visibility through the open source route but for a codec then you have essentially no chance to make any money off of it, ever. have to choose one, but once you do the former there's no going back.
There is the indirect route.. it might make good CV fodder for a better day job. For that, better known trumps.
to make money, I would say keep it closed and patent everything possible, if you have an identified niche where you could license it. that might be what tilman liebchen did with mpeg als.
Varies by country, but for most you have to file either before first release or within 1 year.
Since the command-line encoder/decoder can already be used with wine, what's sorely missing is playback support. How are you going to provide that for the many music players out there, while keeping the source closed? It just won't work. You won't spend time to write a plugin for every music player, and no maintainer is ever going to accept a binary blob in their project anyway.
Yes, indeed. Thomas you shouldn't have become so agitated because this. I think skamp is hinting at a very important point when it comes to a Linux port. I think it's the breaking point of Linux support in TAK.
Encoding and decoding is one thing, and I'm sure you're going to release binaries for that sooner or later. But playback is a lot different, it's a lot more work.
First of all packet maintainers of GNU/Linux distributions will not include this "binary blob", because they are not allowed to. Legally. Period. Consider TAK to be not supportable by many distributions and therefore unaccessible for many users. So providing playback support for Linux at least, would mean for you that you must provide the TAK library by yourself. Downloadable from your website or some offload site you have access to. Essentially this is the same for the simple binary encoders and decoders. Not much different from the situation as it is now for Windows users. I guess some TAK enthusiasts using Linux or Mac OS X are ok with that, but I doubt it will help promote the codec among new users of those platforms. Anyway that's not the real big issue I see. There aren't that many Linux and Mac users on the world anyway...
The far more greater implication for you and your workload when you are actually going to support non-Windows platforms 100% while keeping TAK as closed as it is now is that you will become the maintainer for not only the decoder and encoder binaries but also the playback library. This means whenever there's a major change in the Linux or Mac OS X system libraries that your TAK libirary depend on, you will have to spent some time updating your TAK library without adding any new features.
IMHO it's a big waste of your valuable time. I'm sure you rather want to solely improve the codecs algorithms instead of having to split up your time and also do work on maintaining up-to-date binaries for various other platforms. Also whenever you want to "ship" a new version of TAK (with actual new features that is), you will feel the urge to release it for all platforms at once. And if you don't there will surely come the nagging from non-Windows users...
I don't want to tell you what to do. I just thought by reading your responses and seeing what direction this thread has taken, that this point hasn't been focused on enough when it comes to "TAK for non-Windows platforms".
What I surely don't want to see is that you will try to provide multi-platform support but then get pulled into ever demanding time-consuming support hasslements.
I wish you'd clearly state that there will be no linux support, ever. TAK would be officially a Windows-only codec, and it would put an end to all the drama.
Yeah.
I can't see how keeping TAK closed as it is will lead to a beneficial outcome for both the non-Windows users and you, Thomas. It's just not realistic. It's kind of sad, and I'm not saying that you must open it up so that it spreads to other software and hardware players. You will know best if it's really what you want. Spreading TAK to other platforms, is it really necessary to reach your personal goals? If the answer is yes, can you still reach these goals or will the extra work in fact push you further away from them?
Well, it's not just this way for TAK, I think it's a common problem for every software project that started on one platform is about to become multi-platform. For fairly big projects it costs at least one additional coder for each platform so that new version can be released timely for all platforms.
Or am I wrong? Does anyone know of a closed source project that is able to provide multi-platform binaries in time without putting quite some amounts of manpower into it?
After reading this:
Since the command-line encoder/decoder can already be used with wine, what's sorely missing is playback support. How are you going to provide that for the many music players out there, while keeping the source closed? It just won't work. You won't spend time to write a plugin for every music player, and no maintainer is ever going to accept a binary blob in their project anyway.
Yes, indeed. Thomas you shouldn't have become so agitated because this. I think skamp is hinting at a very important point when it comes to a Linux port. I think it's the breaking point of Linux support in TAK.
And as some time has passed, i see that i may have reacted too harsh:
Aren't you a bit nitpicking? Would my statement only be honest if i added a list of possible player plugins and their exact release dates? Do you expect me to be a lot more precise than the vendors of paid software?
Not at all. I'm trying to make a point. I think linux support without opening up the sources is not realistic.
Sorry Skamp! Affected by the context of some other posts in this thread, i have interpreted your post in an inadequate way.
another misconception here is that ideas can be stolen. they cannot, they're not like bars of gold. no amount of 'imaginary property' laws can usurp natural law. you either have to keep it secret (until someone else picks the same idea out of the ether), or set it free with no attachments.
Maybe it's more like this (i am by no means a philosopher!):They are bars of gold (or possibilities), which exist in inexhaustible amounts and are basically freely available for everyone, but are buried very deep. If one takes them away from someone else, he is stealing his work (exvacating), but not the gold.
to make money, I would say keep it closed and patent everything possible, if you have an identified niche where you could license it. that might be what tilman liebchen did with mpeg als.
Money would be nice, but i haven't done it for money and furthermore would see only very little chance to sell it. The only possibility would be something like this: A big company want's a lossless codec, a bit of money doesn't matter. No, they want it fast and they want something which contains patentable technology, not because they are interested into the technology, but because they want a format they own. Unfortunately i am mainly against software patents... I can imagine some situations, were they may be justified, but mostly they do nothing good.
I forgot to quote DonP:
There is the indirect route.. it might make good CV fodder for a better day job. For that, better known trumps.
That's quite close to what i would like to achieve.
Or am I wrong?
I don't think so.
I have to agree to most if not all of what you said...
I have to find a good solution for my (and also the users) dilemma. But i will not repeat the (open source announcement) mistake from some years ago and present a plan too early.
The only possibility would be something like this: A big company want's a lossless codec, a bit of money doesn't matter. No, they want it fast and they want something which contains patentable technology, not because they are interested into the technology, but because they want a format they own. Unfortunately i am mainly against software patents... I can imagine some situations, were they may be justified, but mostly they do nothing good.
Maybe ask the Nero AG if they need today's most efficient lossless codec? Or if they need a new developer?
1) Couldn't you at least release the source code of the decoder?
No. Unfortunately the source of the decoder would already reveal most of the codec's technology. Most of it's compression power could then be achieved by adopting the code. Only my very fast encoding algorithms would be kept secret.
NO. If it gets popular and remains closed, it will get reversed, as happened to WMA, WMV, NTFS, World .DOC and other things.
My current preference is a later fusion with FLAC
Completely different new codec under old name ? (http://www.hydrogenaudio.org/forums/style_images/ip.boardpr/folder_post_icons/icon9.gif)
I can't bitch about FlaCUDA being restricted to Windows right now, I can only hope Gregory (or yet another third-party) will port it over to my platform of choice (linux).
For the record, it now runs under linux (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=77540&view=findpost&p=677626) just fine, and from Gregory's lightning fast response I assume it didn't work before only because he hadn't tested it (and because linux users like me didn't bother to give it a try until now). FlaCuda was multi-platform from the start.
I'm impressed by TAK but have stopped using it in favor of FLAC (with or without lossyWav) so I can use (play) the files on hardware platforms too.
It's good you brought this up and were honest about regretting your original promise. Now that's out of the way, it's up to you (TBeck) what to do. Take the time you need to evaluate your options.
My 2ct are: there was a momentum for a more efficient lossless codec when you started with it all, but it might disappear over time. Processing power and storage capacity keep increasing. Although, as TBeck says, it is better to use the best algorithms, this is not always what happens in the end.
Newest Flac versions (e.g.): "Flac 1.5 (TAK Powered)"
May I suggest FLACaTAK (pronounced "Flac attack")
If you don't think so, then you should probably make the mob happy.
This "mob" has really only been just a couple of people; at least those who have been vocal about it in a somewhat less than respectful way.
If his optmizations are merged with FLAC, then he would benefit all FLAC users.
The best scenario IMHO would be an jump in FLACs version (be it 1.5, 2.0, etc) with a press release citing the substantial speed/compression improvements and saying that "All this is thanks to TBeck, that ported several algorithms he has developed in TAK over the years" or "Thanks to TBeck's work, we bring to you a new release of FLAC with the following improvements...", or something like that. It would be probably picked by many websites, unlike the TAK 2.0 release, for example.
But FLAC must remain backward/foward compatible. I agree with DOS386 that an different codec with the same name is confusing. The tak files must at least continue to have an different file extension, if tak's encoder/decoder is going to be distributed together with FLAC libs. If tak ever get to be muxed in .ogg files, this would also generate confusion (unlike mka for example, people expect only vorbis and rarelly flac inside .ogg files).
That is my 2cents.
another misconception here is that ideas can be stolen. they cannot, they're not like bars of gold. no amount of 'imaginary property' laws can usurp natural law. you either have to keep it secret (until someone else picks the same idea out of the ether), or set it free with no attachments.
In the US patent climate that seems a very dangerous thing to suggest, Josh. Not to seem like a GPL Groupie, but setting it free with no strings attached may very well result in someone else taking your idea and running with it (to the USPTO). Sure, supposedly ideas cannot be patented, but it's getting very close to that with those process and business methods etc. patents that are given out these days. Anyway, on to my main post.
To TBeck:
While I'm a bit late in adding my voice to the chorus, I just wanted to congratulate you on the work you've done so far. From an April fool's joke to a format that's technically more efficient and faster than most others codecs is no small feat. As such, it's a shame that there doesn't seem to be much place for your codec in it, at least not in the foreseeable future. Similarly, some of the other reactions here have been rather unpleasant in that they seem to be somewhat oblivious to the fact that you "also" have a say in deciding what happens to TAK. Unfortunately, intellectual property rights are a rather contested issue, and people sitting on the extremes are generally the most vocal. ("we have a right to the fruits of your labor" is something Marx strongly disagreed with when factory owners said it, though it's unclear to me what he would think of the fact that the proletariat is now making the same claim. )
In any case, my own opinion is that you should just do what you think you would enjoy most: either decide to just keep on going (regardless of adoption rates) and just have fun trying to make the most of the codec, trying different things to see what works best, or merge with FLAC when you feel you don't want to put more time into doing so. (Although I obviously favor either of these options over you just stopping altogether.)
(Full disclosure: I currently am a MAC -c4000 user who doesn't really see a reason to recode his library.)
another misconception here is that ideas can be stolen. they cannot, they're not like bars of gold. no amount of 'imaginary property' laws can usurp natural law. you either have to keep it secret (until someone else picks the same idea out of the ether), or set it free with no attachments.
In the US patent climate that seems a very dangerous thing to suggest, Josh. Not to seem like a GPL Groupie, but setting it free with no strings attached may very well result in someone else taking your idea and running with it (to the USPTO). Sure, supposedly ideas cannot be patented, but it's getting very close to that with those process and business methods etc. patents that are given out these days. Anyway, on to my main post.
Wrong, only ideas can be patented. More preciselly, only specific types of ideas, listed in the country's patent law, can be patented. This list has been growing over the last 3 centuries, but I hope it starts to shrink again (software patents/bussiness methods patents are seriouslly harmfull, for example).
Also, it is expansive to make and mantain an patent. In many countries you have to pay an anual fee for the maintence of the patent, in addtion to all the costs to first make the patent. If you don't plan on making money on the idea, or don't need nuclear arms for an cold war, it is generally not an good idea to fill a patent. If someone try to steal your idea using the patent system (otherwise it is "only" a copy), and USPTO aproves it, you still have the
prior art defense, that should work.
An usefull rule of the thumb when talking about the many types of "intelectual proprierty" laws: if you know something about copyright laws, for example, you can assume that in patent laws, or trademark laws, or trade secrets laws this thing will be different. In most cases you will be right, as those laws have different origins, talk about different things and evolved diferently.
I will not comment on Marx because that would be going waaay off topic. PM-me in case you are interested.
If someone try to steal your idea using the patent system (otherwise it is "only" a copy), and USPTO aproves it, you still have the prior art defense, that should work.
If you have enough money for the process.
What seems obviously overlooked is the time-consuming process of securing intellectual property. TAK has demonstrated that it has special sauce under the hood, so why shouldn't proper credit go to its creator? From that perspective the entire open/closed question of TAK is premature at this time. I think a logical step would be helping the author with the legal and license issues. For example, who even knows which geographical jurisdictions certain licenses cover?
Another thing that seemed obvious is that new hardware player development is ongoing. I really do not see TAK "losing" here despite its current lack of hardware support. (slightly OT: When I recall Vorbis going v1.0 the adoption was slow because it was not being "added-in" to existing HW, usually a new player supported it from the start.) I don't see the need for a new format to support old devices when there's a trend of faster and more extensible DAP's becoming available, and it seems obvious lossless on portable players is not mainstream enough at this time to be high priority.
Being this is HA, why not suggest things like setting up a PayPal (or whatever equivalent) donation account for TAK development, licensing, legal filing fees and blah blah blah. Although this suggestion seems obvious I don't recall it being stated prominently.
Being this is HA, why not suggest things like setting up a PayPal (or whatever equivalent) donation account for TAK development, licensing, legal filing fees and blah blah blah.
Count me in for €100 for a copy of the source code under a FSF-approved license (http://www.gnu.org/licenses/license-list.html#GPLCompatibleLicenses).
First of all packet maintainers of GNU/Linux distributions will not include this "binary blob", because they are not allowed to. Legally. Period.
just wanted to point out, that this is not true. most distributions include binary blobs. usually for hardware support.
you can find a list of completely free distros here at the fsf (http://www.gnu.org/distros/free-distros.html). the most popular ones are not among them. but i still agree that as a closed format tak will have a very difficult time in linux.
as a former tak user (while i was still on windows) and linux user i would really welcome TAK to go open source. but even apart from that i think it is the right way to go. if recognition for your work is important to you thomas, think about this: your "fame" is proportional to the success of your project. whether it is open source or not. open sourcing will definitely boost software and hardware adoption which would make it more successful, maybe it will even attract contributors to help you improve the codec.
merging tak's features into flac would be the coolest thing of them all: combining tak's performance with flacs popularity and hardware support. maybe in a new major version flac v2. but, i've no idea if that is even technically possible.
tom and josh joining forces to work on the greatest and best lossless codec!
i don't know about you, but to me this sounds like a dream come true
thanks for answering our questions in this thread thomas
please let us know where things are going.
TBeck, can you provide an estimate on the amount of functions or speed/compression increase from TAK that are implementable into FLAC without breaking backwards compatibility? If it still doesn't make FLAC improved this way as efficient as TAK 2.0 then you probably shouldn't hesitate with the merger, as you will still retain your niche. :)
Anyway, I actually came here because I figured I could reencode my lossless collection into TAK, but knowing there is no out-of-the-box Mac/Linux and potential Rockbox support I guess I shouldn't. Many thanks for doing this, though; I've been following TAK's development in its early days, and you've been doing a terrific job improving both speed and compression ratio while nearly no other lossless codec progressed in the mean time. It's worth a lot no matter what TAK's future will be.
Just don't let your work fade into obscurity, that will be the worst scenario.
One thing worth considering with licensing is selling license exceptions. When a copyright holder releases their source code under a copyleft license such as GPL they are not giving away their rights to offer the same code to different parties under different terms. So if TAK was released under GPL (or similar copyleft) it would become easily useable in free software projects and would almost certainly eventually be packaged for Debian, Fedora etc and so be available in all their numerous derivatives. But the author may still sell license exceptions. For example a company might want to use the code in a proprietary project, this may not be possible with GPL'd code. But the copyright holder i.e. the author of TAK, when licensing the code as GPL did not give up his copyright or his right to offer the same code to others in a different way, so he can offer the code to the interested company and sell them a license exception (actually he can give them the exception but selling is more likely in this context). Nobody else can do this, only the copyright owner. There is a very interesting article on the subject by none other than RMS Selling Exceptions (http://www.fsf.org/blogs/rms/selling-exceptions)
Before I posted I'd thought that this might have been mentioned before but a search showed only the idea of dual licensing, which is slightly different and anyway the proponent suggested dual GPL/LGPL which makes no sense afaik.
My 2 Pfennigs:
What's the purpose of TAK? What will you (i.e. TBeck) use it for? Well, there is a small probability that someone will call you up, saying "I want this, what's your price?" but I doubt it. The lossless market isn't that big, and the improvement over the peers, while noteworthy, is not revolutionary (I mean, 10 percent up in HD space or 20 percent gain in speed doesn't save lots of money!) The niche for "closed" formats -- for DRM? -- is miniscule, and without being pushed by Windows and Apple (who already have their own formats), TAK will never be widespread as a closed format. Probably not as an open either (except as merged into ... say, FLAC as someone suggested), because honestly, who does really desperately need TAK?, but at least with a fairer chance as open.
This is not unrelated to the Linux support; lots of devices run Linux (e.g. Squeezebox, Android). But even more important: lots of devices cannot use your Windows binaries, and will stick to formats with sufficiently open specification. That does not necessarily mean that your encoder/decoder has to be open source.
Publication is of course irreversible, so do not open up until you are fairly sure that this is what you want. And if you want to finish what you started before handing it over to someone else, fine. I don't need your source code
(I appreciate its existence, as there isn't much other development of lossless audio these days. But the lack of support and openness keeps me safely in FLACland, where I know I can still play my TBs of music the day Bill Gates provides the next OS update "which may render 3rd party programs unusable" and by coincidence (or not) Jcoals is hit by the train.)
@TBeck
There's a way I think you could avoid to make TAK code to be just stolen, and then TAK just forgotten . Just try to promote the format as much as possible, when the source code is released.
Use advertisements. Talk of the format and advantages to a lot of people. Try to make it adopted by hardware manufacturers. Now if someone re-use your code, his new format won't be popular, because everyone would have heard or already adopted TAK. Unless he improves TAK compression algorithms , and keep the improvements for himself.
If I might elucidate as a mere fan of lossless compression in general, the fear that TAK's code to be stolen is not as much the issue as the credit for work done to date. Thomas has been a a "hobbiest" of lossless compression and his personal project has been going for quite some time before he decided to mention to the public that his formula was quite competitive to existing ones.
Since going public TAK has already used in at least one commercial software iVST installation software, has a DShow filter and other public software that allow it to be used in Winamp and FB2K, and there are likely other implementations that I have managed to overlook.
In short: it works. And damn good too
The question is: What can YOU (and I) do to help TAK and the author achieve beneficial success to the user and the author alike?
The major issue I can see is with licensing. It seems to me that licensing is tricky and involves a lot of research. Remember that TAK was a hobby before legal and patent aspects started to infringe. Thomas should be granted credit for his contributions if the tech is, in fact, original and can be described and patented under existing regulations. This requires legal expertise and confidentiality, not to mention the time and money to apply for such licenses. A sole person doing this as a hobby can not examine these aspects without stalling development.
Please bear in mind I'm not an expert with legal and business aspects of audio tech, I just want a fair outcome for people who have done their homework. Personally, I think TAK is a very nice addition to the lossless compression community if it's only to keep the other lossless audio developers on their toes.
I'd like to thank Thomas for bringing TAK to the public and offering it to be used essentially for free and support of the philosophy of better inventions. Thank you!