The latest draft, published today, is a landmark. You can't possibly imagine the number of hours of hard thinking and nasty wrangling that have gone into producing it. My personal take is that it's about done and it's good enough and we're not going to end up with anything better. Warning: long and boring, but I think important.
To start with, it's really important to understand that these issues are not simple. I think that most responsible members of the Web communities know what effect we'd like to achieve, but the legal and historical backdrop is insanely complex and achieving the effect you want to is far from easy.
Royalty-Free is the Way to Go · The Web is unique and special in that anyone can read up on HTTP or HTML or XML or whatever and start writing software to process them, and not have to worry about paying anyone a royalty. Lots of other areas of technology just aren't like this. If you want to build a DVD player, or a network router, or a VLSI test rig, or many many other things, you have to cough up some serious money for patent rights first. That's one reason why there is lots of Open Source web softwre, but no Open Source DVD players. And, it's one of the nice things about working in Web space.
VoiceXML is the Villain · The issue that really forced the W3C to think about this was VoiceXML. This is an XML tagset for spoken dialogues, i.e. you say to the vending machine "Two croissants please", the machine says "That'll be $4.82, how would you like to pay", you say "Visa", the machine says "Insert your Visa card under the flashing light please", etc.
VoiceXML was originally cooked up by the "VoiceXML Consortium", which includes several well-known big names in the telecom and cellular space. And several of them had patents which they claim covered it so you wouldn't be able to use it without a license.
The VoiceXML consortium was quite happy to sign this over to the W3C. To be honest, I fully believe that one or more of them had dollar signs gleaming in their eyes, and biz dev execs cackling in the back rooms, because they were going to get a W3C stamp on their patented technology and put up a tollbooth beside the road, and anyone who wanted to voice-enable the Web could just send a cheque.
Concerns were raised and the W3C formed a Patent Policy advisory group, which initially was made up almost entirely of representatives from big companies with big patent portfolios. Not surprisingly, their first draft said that it would be OK for the W3C to standardize on patent-encumbered technologies. Surprisingly, nobody noticed this gestating policy until the comment period was almost over. Not surprisingly, the proverbial shit hit the proverbial fan.
The group invited in some representatives of the Open Source community and went back to work, and the latest draft is the outcome.
The Draft · This draft shows how much trouble and work can go into trying to achieve a basically reasonable effect. It didn't manage to completely rule out non-royalty-free W3C product, but they made it pretty damn unlikely. Once you agree you want royalty-free, then you have to worry about a lot of important details:
The result of all this is contorted rules like the following:
4.3. Joining an Already Established Working Group
Participants who join a Working Group more than 90 days after the publication of the first public Working Draft must exclude Essential Claims covered in the latest Working Draft published within 90 days after the first public Working Draft at the later of 150 days after the publication of the first public Working Draft or upon joining the Working Group.
Software Patents in General · The elephant in the room that nobody's talking about is the widely-held conviction that a large proportion of software patents are inherently bogus, and fail the basic test of "obviousness" that's supposed to constrain the granting of patents.
I personally believe this, and have read quite a few software patents, and frankly think most them are complete bullshit. I have yet to encounter a software patent where, if someone had told me "I want to write a program to do XXX - could you do that" I'd have to say "No, I wouldn't know how to do that."
I don't believe, as for example Richard Stallman does, that software patents are inherently self-contradictory and wrong; there are algorithms, for example in the cryptography space, where the math is sufficiently hairy and deep as to represent a genuine "invention", an unforeseeable intellectual leap forward that might qualify under the spirit of the patent legislation.
Another example: I suspect that if Ken Thomson and Dennis Ritchie had at
the birth of Unix applied for a patent on "a method for creating processes
based on a 'forked' copy of an existing process", it might have been granted
and might have been deserved.
fork() primitive is highly non-obvious and you don't
realize what a great idea it is until you've had a
chance to work with it for a while.
The fact that I'm open to the idea in principle should, I think, lend weight to my opinion that the US PTO has been indulging in serious bad craziness in recent decades, putting their stamp on many so-called inventions that are frankly silly.
Disclosure: I have my name on two pending patent applications for some of the protocols and effects achieved by the Antarctica software. These, of course, are of the highest intellectual quality and represent Real Inventions. At least that's my story and I'm sticking to it.
On the Stupidity of Some Patent Holders · But let's suppose that software patents were in fact solid and respectable, does that mean it would be a good idea for the W3C start blessing their application to basic Web technology? No, and to illustrate why, here's a remark I heard from an IBM-employed lawyer: "You know, we realize over a billion dollars a year in patent licensing revenues."
BZZZZZZZZZZZZZZZZT! And, you moron (I didn't say) how many dollars in revenue is IBM pulling in because the Internet and Web are here and are turning the world inside out, based largely on Open Source software and unencumbered protocols!?!?!? I'm delighted to report, by the way, that IBM has apparently realized this and played a very constructive role in the patent policy work.
The Bottom Line · I think this policy draft is important in that it's basically sound and well-thought-through. I think the real significance though is psychological - the process in the W3C that got us here has established that any attempt to put up a patent-powered tollbooth on the Net's software will be met with fierce popular and legal resistence, to the extent that it may be self-defeating.
And in closing, I think we all owe a tip of the hat to the W3C Patent Policy Working Group for fighting through this and getting us to a basically reasonable end-point; in particular I've been impressed at the work of Danny Weitzner, who also edited this draft. Thanks, guys.