And of course I'm planning to describe the algorithms and the exact file format in a detailed and public specification, which should be accurate enough to allow anyone to write their own FLIF implementation.
And of course I'm planning to describe the algorithms and the exact file format in a detailed and public specification, which should be accurate enough to allow anyone to write their own FLIF implementation.
Game development projects, in particular, will avoid it, as will almost anything that wants to publish an application to the Apple App Store or Google Play.
You may wish to consider whether adoption of a standard implementation is more important than other goals you may have considered in choosing a license.
"Once the format is finalized, the next logical step would be to make a library version of it, which will be most probably get licensed under the LGPL v3+"
Basically anything copyleft is going to have a really big dropoff in use compared to say BSD, MIT or Apache 2.0.
A key attribute for companies in a heavy competitive market is that they have to use every tool available to get the best product out in least amount of development time. If a company decided to avoid a license just because of religious reasons, 10 other companies will jump into its place and out compete them.
A good example of this is when we see companies reaches out to developers directly to get permission, as the time to do that is worth compared to having developers spend time on unnecessary code. Large AAA games can easily have a long credit list of every license available that can be used in a proprietary product, including LGPLv3.
Yes, some platform holders will use it themselves, but that is not their preference. Ultimately, image encoding and decoding is a crowded market with many alternatives and I strongly believe that a copyleft-licensed component will always be passed over in favor of alternatives unless it is overwheingly compelling and a viable option.
Some companies explicitly prphibit the use of any copyleft licensed software in their development.
Certain conditions of the GPL and LGPL are impossible to fulfill on some platforms.
With that in mind, I feel like my assertions remain reasonable.
Cool work in any case!
One thing the page could use is "how much processing power does this use, compared to other things?"
Without a BSD/MIT/ASL2 license as an option, you'll never see this in Internet Explorer or Chrome. Probably not even Firefox.
[1] Companies license patents in bulk from other companies for their own use all the time. They don't have the right to sublicense those patents to others; they're just protected against any lawsuits relevant to the use of the ideas in those patents. Yet GPLv3 requires that they provide a free license to any patent that they have a license to that might be required to use GPLv3 source code. So it's requiring them to do something they legally can't do. Selecting GPLv3 means that no large company will ever touch it as a result.
The only thing (L)GPLv3 patent clause obliges propagators to do (ie not mere users; not even mere modifiers) is to grant their users licenses to applicable patents which they own. The permissive Apache license v2 demands from contributors the very same. It is the software patents that should be "banished", not freedom preserving licenses.
TL;DR: (L)GPLv3 prevents patent trolling through free software.
Disclaimer: This is not to be treated as legal advice.
Doesn't matter what you or I think of software patents (yes, they should be banned). Doesn't matter what you or I think GPLv3 says.
The lawyers at big companies see it as a problem, so it's a problem. End of discussion. No drama required; it's just the fact that big companies avoid using anything cursed with GPLv3.
Note that what you've argued before was very different from what you're saying now. You've narrowed the scope of the discussion (leaving out LGPL), but also its very nature ("legally impossible [1]", eh?).
Anyway, companies do use software licensed under both licenses, and even incorporate them into their services - hence the need for AGPL. Maybe others wouldn't see GPLv3 as that much of a problem if people didn't spread FUD about its supposed "curse" ("no drama required" but you couldn't help it, huh?). And if they didn't defend harmful practices, like Linus does tivoisation.
But mostly what companies avoid is copyleft, because it mandates reciprocity and prevents leeching the community. For projects such as these, LGPL is an acceptable compromise. The only valid argument against it is that apps under incompatible licenses will not be able to use it where dynamic linking is barred, and such is the requirement for apps in the Apple's store. However, in this particular case, that wouldn't be a problem either if the platform itself provided a decoder, like iOS does for PNGs.
I have read the license. I'm like that.
And it still doesn't matter.
It's what the lawyers think. And they say it's verboten.
* You are lying that you read the license.
* You were lying about what the license says.
* You really dont want to admit that you have misunderstood the license.
Either way, you were wrong then and you are wrong now about what the lawyers think. Speaking of which, there are only a few possibilities here as well, only these are not mutually exclusive:
* You are intentionally dishonest because you have an agenda
* You are dishonest just to cover your behind
* You are genuinely careless about what you say
Even if we change the word 'think' for 'say', it's still a gross overgeneralization.
So all things considered, in the best case scenario, you refuse to admit when you are wrong, and will continue overgeneralizing. Forgive me, but it's really not worth the effort arguing under these circumstances. If you wanted to continue, you would have to make some concessions, but I doubt you will, so in all probability: Goodbye.
I was reporting what I was told by corporate lawyers. My own reading of the patent section does happen to side with the lawyers' reading: That if you distribute an app that's protected by a patent you own a license to, that you need to arrange a sublicense for all users of that software. Maybe not technically "impossible," but I didn't count "spending millions of dollars to fix the problem" among the likely corporate responses when I said "impossible." Especially when most GPLv3 code can be written from scratch for less than the cost to license patents.
Someone alleged Blizzard uses it; fine, their lawyers either disagree, weren't consulted, or are being ignored, but Blizzard doesn't make Chrome, Firefox, or Internet Explorer, so the point is moot if you care about web adoption, which would make the format relevant to anyone but a game developer.
What matters is what the lawyers for the big companies that control Chrome and IE won't let GPLv3 code into the code base. Many other big company lawyers take the same position (probably all companies above some size threshold), and that's all I've been alleging from the start. Criticize my delivery all you want, but that's what I was trying to say.
My agenda is to get the developers to change to a license that could actually be adopted into a web standard. Since you're refusing to actually read what I'm saying, I agree: Goodbye.
Companies wouldn't adopt things under "GPLv3", but they wouldn't a permit GPLv2 either. Or LPGL. Or Apache 2. Or MIT, or BSD, or any license. They permit nothing short of contributors assigning them copyright and the patents, just them (see eg. Webkit's and Chromium's copyright notices and CLAs). And yet, libpng is under a license. So yeah, I agree they'd write their own library - out of their selfishness. Let them.
With the "adopting a web standard" thing you're attempting to further move goal posts. But you fail, and not because your implication that standard bodies would accept permissive licenses is wrong - which it is, because they're exclusively public domain + patent clause (oh and the people building browsers still contribute somehow). You fail because you're mixing apples and oranges again; programs are not parts of standards. Standards describe file formats, and prescribe behavior of programs that process them. They are not concerned with implementations' licenses.
The spec can become a public domain standard, and all would still be well with the library under (L)GPLv3+. Free software should have the edge.
If they are using LGPLv3 libraries, then presumably their lawyers are OK with it, or they failed to run it past legal. Regardless, it's completely irrelevant to my point what Blizzard uses or doesn't use.
If you want to see a lossless compression format on the Web, you need the format to be picked up by, at a minimum, Google and Microsoft, the owners of the two top browsers in the world today.
Firefox would also be important, but would certainly follow if Google and Microsoft stepped up to support it. So really it only matters what Google and Microsoft think. If it's a license that they can freely use, then there's a chance it becomes a new Web standard. If it's only available LGPLv3, then Microsoft and Google would need to buy LGPLv3 exceptions in order to use it. A much bigger barrier to entry, and at least Google might object based on the concept that Web standards should be open. Mozilla would certainly resist if they weren't open -- though they eventually caved on H.264.
If you want to see adoption where it counts, you need cooperation from the companies behind the browsers with the market share.
No one is expecting Starcraft 2 to be the next big Web plug-in, so it's irrelevant to the product that they have LGPLv3 code in it (if they do). And if they do, Blizzard better hope that their lawyers are right in their interpretation of GPLv3, and that no one sues them hoping to take home a share of the profits you've pointed out that they are rolling in by forcing them to buy a license to get around the GPLv3 restrictions that they may or may not be violating. That's an expensive lawsuit even if they win.
That's only guaranteed if you obtain copyright assignment from any contributors now. Otherwise, re-licensing will be a massive headache as you'll need to track down all the copyright holders for permission. Some might refuse or have have dropped off the map.
This seems so... avoidable. Maybe the FLIF format could include a version number declaration?
If they do it for pre-finalization versions, then they (and other implementations) will have to keep supporting them, which probably doesn't make sense considering how small the corpus of images in this format likely is.