How non-free do we consider Flash to be? The Gnash player appears to be making good progress. Would it be acceptable to permit useful Flash files which work in Gnash and don't require non-free codecs to be uploaded?
I did not see any issues with patents mentioned in the relevant Wikipedia article. The old Macromedia Flash website lists a US patent on "creating gradient fills", but that seems so bizarre as to pose no real threat.
(Let's keep this separate, for now, from the question when a format like Flash would be appropriate, content-wise. I'd like to fully understand the "freeness" first.)
On 3/6/07, Erik Moeller erik@wikimedia.org wrote:
How non-free do we consider Flash to be? The Gnash player appears to be making good progress. Would it be acceptable to permit useful Flash files which work in Gnash and don't require non-free codecs to be uploaded?
I did not see any issues with patents mentioned in the relevant Wikipedia article. The old Macromedia Flash website lists a US patent on "creating gradient fills", but that seems so bizarre as to pose no real threat.
For what application? I see that you say 'don't require non-free codecs', but it's not clear that you're aware of the status of codecs in flash:
For video support the flash plugin contains a H.263 video and a MP3 audio codec.
It's not possible to write MP3 software which does not infringe 5,105,463 among several other patents (http://news.bbc.co.uk/2/hi/business/6388273.stm Weee!).
H.263 is covered by a pool of patents similar to MPEG-4.
Flash player 8 and above also includes the VP6 video codec. On2 has patents they license for VP6 use, but overall they are pretty cool company though, so I'm not fully sure where that stands. I expect that requiring flash player 8 would leave Java as a more compatible web toy solution.
Unlike the Java video player implementations the codec isn't actually 'written in flash'. Actionscript is a poor match for doing lots of data manipulation. I understand that someone nearly has a Vorbis codec working actually written in actionscript, but it's too slow to be useful.
So for video, flash does not look good. The use of cortado in Java would be much better. (http://www.flumotion.net/cortado/). Success rate of the inline Java audio player hasn't been bad, and cortado should be even more compatible with older JVMs.
A bigger concern I have with the freedom of flash doesn't have anything to do with patents: There are no free authoring tools, other than some hacks that are used for dynamic authoring. I think this is a pretty big killer. Nothing we need for authoring today requires proprietary software.
If we're looking to allow users to submit flash there another freedom related issue:
Since the toolchain is proprietary and built around desktop usage we'll be stuck with the 'opaque object' problem. In short, for downloadable turing complete software (Java, Flash, JS) it can be impossible to fully understand what the software will do. Perhaps on the 3rd of the month it displays the goastse image, perhaps it exploits the sandbox and steals your data. In cases where we have the source, it's possible to require the software be simple enough that no such tricks could be easily inserted. As a result I previously suggested that if we ever accept user submitted java that we have people submit source and we compile it on our side. Because the flash authoring tools are proprietary this option is closed to us.
These are just some quick points, a little more knowledge of the intended application would help me drill in and look for freedom related limitations.
I hope that I'll be given the opportunity to comment on the non-freedom related aspects, since I think there are some significant issues. Again, knowing the application is critical.
I'm pretty impressed by the Cortado demo. I'm not sure how well it competes with Flash video on lower-end machines (i'm at a high-end machine right now), but it seems like a really good alternative for presenting in-line video on Wikimedia projects. One of the main hassles for incorporating videos on Wikimedia projects (aside from the copyright and bandwidth issues) always seemed to be the lack of widely available Theora support. With everyone being accustomed by YouTube working right out of the box it seems a bit weird that you ask people to download a codec and install it, or use an alternative player. This works right out of the box, in Firefox at least. I had to download the J2RE on IE 6.
If we could use this in some way it would be a great addition to the projects.
-- Hay Kranen / [[User:Husky]]
On 3/6/07, Gregory Maxwell gmaxwell@gmail.com wrote:
On 3/6/07, Erik Moeller erik@wikimedia.org wrote:
How non-free do we consider Flash to be? The Gnash player appears to be making good progress. Would it be acceptable to permit useful Flash files which work in Gnash and don't require non-free codecs to be uploaded?
I did not see any issues with patents mentioned in the relevant Wikipedia article. The old Macromedia Flash website lists a US patent on "creating gradient fills", but that seems so bizarre as to pose no real threat.
For what application? I see that you say 'don't require non-free codecs', but it's not clear that you're aware of the status of codecs in flash:
For video support the flash plugin contains a H.263 video and a MP3 audio codec.
It's not possible to write MP3 software which does not infringe 5,105,463 among several other patents (http://news.bbc.co.uk/2/hi/business/6388273.stm Weee!).
H.263 is covered by a pool of patents similar to MPEG-4.
Flash player 8 and above also includes the VP6 video codec. On2 has patents they license for VP6 use, but overall they are pretty cool company though, so I'm not fully sure where that stands. I expect that requiring flash player 8 would leave Java as a more compatible web toy solution.
Unlike the Java video player implementations the codec isn't actually 'written in flash'. Actionscript is a poor match for doing lots of data manipulation. I understand that someone nearly has a Vorbis codec working actually written in actionscript, but it's too slow to be useful.
So for video, flash does not look good. The use of cortado in Java would be much better. (http://www.flumotion.net/cortado/). Success rate of the inline Java audio player hasn't been bad, and cortado should be even more compatible with older JVMs.
A bigger concern I have with the freedom of flash doesn't have anything to do with patents: There are no free authoring tools, other than some hacks that are used for dynamic authoring. I think this is a pretty big killer. Nothing we need for authoring today requires proprietary software.
If we're looking to allow users to submit flash there another freedom related issue:
Since the toolchain is proprietary and built around desktop usage we'll be stuck with the 'opaque object' problem. In short, for downloadable turing complete software (Java, Flash, JS) it can be impossible to fully understand what the software will do. Perhaps on the 3rd of the month it displays the goastse image, perhaps it exploits the sandbox and steals your data. In cases where we have the source, it's possible to require the software be simple enough that no such tricks could be easily inserted. As a result I previously suggested that if we ever accept user submitted java that we have people submit source and we compile it on our side. Because the flash authoring tools are proprietary this option is closed to us.
These are just some quick points, a little more knowledge of the intended application would help me drill in and look for freedom related limitations.
I hope that I'll be given the opportunity to comment on the non-freedom related aspects, since I think there are some significant issues. Again, knowing the application is critical.
foundation-l mailing list foundation-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/foundation-l
On 3/6/07, Husky huskyr@gmail.com wrote:
I'm pretty impressed by the Cortado demo. I'm not sure how well it competes with Flash video on lower-end machines (i'm at a high-end machine right now), but it seems like a really good alternative for presenting in-line video on Wikimedia projects. One of the main hassles for incorporating videos on Wikimedia projects (aside from the copyright and bandwidth issues) always seemed to be the lack of widely available Theora support. With everyone being accustomed by YouTube working right out of the box it seems a bit weird that you ask people to download a codec and install it, or use an alternative player. This works right out of the box, in Firefox at least. I had to download the J2RE on IE 6.
If we could use this in some way it would be a great addition to the projects.
*Sigh* We had a SOC project last year to provide integrated mediawiki support for this, but it appears to have fizzled.
I threw up an audio only version (using jorbis) a while back to demonstrate that it could be done for the non-believers. It's used many tens of thousands of times a week. The only complaint I get about it these days is due to a longstanding image server problem which causes us to send zero byte files which get stuck in squid, rendering some audio files useless.
I was hoping that my quick hack would inspire someone to do it right and include video support, but it hasn't yet. It would be easy enough to do the same kind of hack for video, but we really should get proper support.
As far as performance goes, it will run great on any system which as good as the system requirements for modern versions of flash. :)
Depending on who's numbers you believe Java penetration is someplace between very sightly less than flash and substantially less but still a clear majority. I suspect it depends on what sort of user's you're polling.. Java being more commonly found on computers used for business, while flash is almost excursively used for entertainment.
The last weeks data of 143,530 unique IPs shows that 79% of the IPs hitting the Java audio player on toolserver have a Java Virtual Machine. However, of those only 78% (61% of the total) have a JRE new enough to use Jorbis. I believe Cortando has a resampler in it to permit it work on the old MSFT JVMs, so it should do better. The number is skewed a bit by the fact that people without java are unlikely to vist again soon since it didn't work, and are somewhat better than my initial numbers, but I've also improved compatibility dramatically since then.
These numbers are consistent with the higher estimates of Java penetration vs flash.
wikimedia-l@lists.wikimedia.org