- From: Maciej Stachowiak <mjs@apple.com>
- Date: Sun, 26 Mar 2006 22:54:33 -0800
- To: Web APIs WG <public-webapi@w3.org>
Some corrections. I think the rest of what I said is right though. On Mar 26, 2006, at 10:47 PM, Web APIs Issue Tracker wrote: > > 1) A technical specification should not attempt to legally require > attempts at implementation to > conform, just on general principle. It would be unprecedented for > the W3C, certainly. Apparently not, it looks like this is in the published DOM Level 3 Events Note, and in DOM Level 2 Events. > 2) Taken to the letter, this would mean that any unknown (and > therefore undocumented) failure to > conform in an attempted implementation would be a copyright violation. > > 3) Taken to the letter, it would mean any not yet complete Java > implementation of this specification > would be a copyright violation. It would have to not use the w3.org > prefix and therefore not implement > even a subset of the specification. This part is not right either, for the same reason: > That seems like a bad idea, especially since DOM Level 2 Events is a > subset and places no restrictions, creating the silly situation > that it is legal to implement DOM Level 2 > Events only, but so long as you add a single L3 method you need to > change your package prefix. > 4) Subset profiles of DOM Level 3 events would not be legally > possible. > > 5) The requirement in the W3C software notice "Notice of any > changes or modifications to the files, > including the date changes were made" are incompatible with the GPL > and the LGPL, which forbid > placing any additional requirements on modification or > redistribution. This would make it a copyright > violation for many popular open source implementations to implement > the spec. At least if they looked > at the spec - it may still be legal to do a clean room > implementation where one person reads the spec > and tells another what to code. Which is crazy. > > I recommend instead that Web API specifications make clear that > anyone is free implement the > interfaces therein, whether correctly or not, but of course may not > make factually incorrect claims of > conformance. > > > >
Received on Monday, 27 March 2006 06:54:52 UTC