> This is quite a good summary: > > http://gstreamer.freedesktop.org/data/doc/gstreamer/head/faq/html/chapter-legal.html
Thank you for that pointer. It gives definitive answers to many questions. Of course, it doesn't ask and answer *all* of the questions we're interested in. It's mostly focused on proprietary binary plugins, which isn't what I'm talking about. I'm talking about GPL or LGPL code that implements algorithms that are (likely) patented in the US and/or other countries. See below. > "We are not prepared to issue and irrevocable, free patent license to > GPL and LGPL software using our patents." > > This means MP3 will never be free software compatible. > ... > The only long-term solution is to lobby Macromedia to include Vorbis > support into a future open Flash standard. No! MP3 will be free software compatible as soon as those patents expire. One of the beautiful things about patents is that they only last 15 to 20 years (as compared to copyrights, which exist for 70 or 90 years or more). Many patents useful to free software have already expired, such as the RSA and Diffie-Hellman patents. The long-term solution will be to do whatever we want, less than 20 years from now. Each of the relevant patents expires at a particular different time. If one or more of the patents expires later than the rest, it may also be possible to write or rewrite an MP3 decoder so that it avoids that particular patented technique. E.g. some of the patents only apply to methods for encoding, so decoders don't violate them. Some other patents may well cover one way to decode, but there may be another way to decode which is less efficient but not patented. (Or we may invent a more efficient way, which is also not patented.) Apple's TrueType font patents were worked around in this way for the FreeType library. The details are here: http://freetype.sourceforge.net/patents.html I don't know of anyone who's tried to do this with an MP3 encoder, or an MP3 decoder, but we shouldn't rule out this path before checking whether it is viable. We should determine the earliest date on which a non-patent-encumbered MP3 decoder could be made. With that information in hand, Gnash (and other projects) can make a much better decision about how we support MP3. > "We do allow non-commercial use of our patents. However, the GPL and > LGPL software allow onward distribution that can easily be non-commercial." > > In other words, the minute we interpret your free software project as > commercial, we'll take legal action. I think their intent is clear: * They distinguish between "commercial" and "non-commercial". * Our software is free to be used either way; we don't require one or the other. Therefore, people who derive revenue from selling or using our software are expected to pay for a patent license. People who derive no revenue from distributing or using our software are NOT expected to pay for a patent license. > For example, the Linux magazine that I edit has a free DVD or CD on the > cover each month. Clearly, the cover disc is part of the value of the > magazine for readers - so is that commercial distribution? If you make money from publishing the magazine (either by selling ads, or by selling the magazine to its readers), it's commercial. Red Hat and other commercial Linux distributors are commercial distributors, too, and would have to pay for the patents. But Debian, which doesn't charge money for distributions, nor sell ads, would not have to pay. Fedora, similarly. FSF, similarly, would not have to pay. That's my theory; shoot more holes in it if you like. > But then you'd be infringing the GPL, because the patent licence almost > certainly won't be GPL compatible. This is why Fluendo have signed the > MP3 deal, then based their system around an LGPL media player (Totem) > and binary plugins. "GPL compatible" is not a yes-no concept. Potential GPL incompatability because of patents differs from potential GPL incompatability related to copyrights, for example. And in GPL3, which is in development, a lot of work is going into expanding GPL compatibility. Even if what we decide to do doesn't work in GPL2, it may work in GPL3. Fluendo chose one (relatively bizarre) path through this thicket. I'll go into more detail on their path after I have a chat with them about some of the details. But this is not the only way to license an MP3 decoder. It's just the way that Fluendo chose, for their own business reasons, to do it. Libmad chose another way, for example. > But then you'd be infringing the GPL, because the patent licence almost > certainly won't be GPL compatible. That's apparently not true. As the GStreamer page you pointed me to says: Q: There are a lot of GPL- or LGPL-licensed libraries that handle media codecs which have patents. Take mad, an mp3 decoding library, as an example. It is licensed under the GPL. In countries where patents are valid, does this invalidate the GPL license for this project? A: The mere existence of a patent which might read on the program does not change anything. However, if a court judgement or other agreement prevents you from distributing libmad under GPL terms, you can not distribute it at all. The GPL and LGPL say (sections 7 and 11): "If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Library at all." Q: So let's say there is a court judgement. Does this mean that the GPL license is invalid for the project everywhere, or only in the countries where it conflicts with the applicable patents? A: The GPL operates on a per-action, not per-program basis. That is, if you are in a country which has software patents, and a court tells you that you cannot distribute (say) libmad in source code form, then you cannot distribute libmad at all. This doesn't affect anyone else. John Gilmore _______________________________________________ Gnash mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnash
