- From: Thierry Michel <tmichel@w3.org>
- Date: Mon, 04 Dec 2006 08:07:35 +0100
- To: Lofton Henderson <lofton@rockynet.com>
- CC: public-webcgm-wg@w3.org
Lofton, see my comments in line with following header. >One question gets to my mind, does WebCGM 2.0 REC supersedes 1.0 ? Hmmm... interesting question. My inclination is "no". Detailed thoughts follow... BACKGROUND. I looked at http://www.w3.org/TR/, and I find only one Recommendation that carries a version > 1.0 and that even mentions the topic (at least in SoTD) -- SMIL 2.1. It supersedes 2.0. It links to 2.0, but oddly, 2.0 cannot be found in the list of Recommendations in /TR/ -- it has disappeared from the list of Recommendations. --> yes because 2.0 was superseded by 2.1. Therefore the comm decided to - remove the link from the TR page. - have the "latest SMIL 2 version": http://www.w3.org/TR/SMIL2/ link to 2.1 (not 2.0). (Note also: the SoTD of the old SMIL 2.0 spec, once you find it, does not announce that it is superseded, obsolete, defunct. Could a PER publication do that -- mark a version as superseded?. TM --> one can not update a Recommendation as it is a frozen document. TM --> My question about WebCGM 2.0 superseding 1.0 is for that reason. If it does than WebCGM 1.0 will disappear from the TR page. I looked in the Process Document. The word "supersede" does not appear. Neither does it appear in pubrules. I found it here: [2] http://www.w3.org/2005/05/tr-versions I don't know what exactly what "supersede" means to W3C. In ISO, the terminology is "supersedes and replaces" (S-R), and it means "[old version] is no longer a standard". So in ISO, version 2 CGM S-R'd version 1 CGM, version 3 S-R'd version 2, version 4 S-R'd version 3. But in the ISO CGM case, each new version incorporates the prior one as an identifiable conformance level. TM --> Right there is no clear mention of that at W3C. ABOUT WEBCGM. It would probably have been good for us to have looked at and discussed this earlier. But if we agree with "no", then it's probably not a big deal. TM --> I would say that the easiest is to say "no". All of that said, IMO the answer should be "no", 2.0 does not supersede 1.0. Why? While it adds to and builds on 1.0, it would create a problem to say that the 2.0 spec superseded the 1.0 spec. The problem is that 2.0 spec, for example, identifies ProfileEd:2.0 content, but does not identify ProfileEd:1.0 content. If 1.0 were no longer a standard, what is the status of 1.0 content and 1.0 implementations? Rather than raise such questions, I think it is cleanest to keep 1.0 updated/corrected with errata (and "Editions", which according to [2] *do* supersede the earlier editions). MORE. We have explicitly talked in the 2.0 spec, in at least one place, that "2.0 conformant viewers that receive 'foo' in 1.0 content shall blah..blah.." This reinforces the perception that 1.0 remains a valid conformance target for content, generators, and interpreters that do not care about 2.0's extensions and additions. MORE. Note that WebCGM 2.0 also has a different shortname: webcgm20. If we explicitly intended "supersede", should we have used the same shortname? Note also that the Abstract says that 2.0 "adds [stuff]... builds upon and extends...", which to me implies that 1.0 is still available for simple jobs. CONCLUSION. 2.0 does *not* supersede 1.0. Corollary: We have to keep 1.0 correct and accurate with errata and possibly edition 3 [Third Release]). Thoughts? TM --> This looks fine. Thierry
Received on Monday, 4 December 2006 07:07:51 UTC