- From: <noah_mendelsohn@us.ibm.com>
- Date: Tue, 4 Sep 2007 18:16:31 -0400
- To: www-tag@w3.org
On 13 August 2007, the TAG discussed [1] my initial draft of a blog entry on version identifiers [2]. In response to comments received then, as well as email comments sent to www-tag, I have prepared a new revision (Latest version always at [3], permalink to this new version [4]). One of the things the TAG clarified on the 13 Aug. telcon is that TAG blog postings will in general represent the opinions of individual TAG members. So, there is not (and in retrospect never was) any formal review process that is (was) needed before this gets posted. Nonetheless, the comments received have been much appreciated, and I hope this draft has been to some degree responsive. Among the suggestions I've tried to address are: * Making clear that the first Good Practice Note is quoted from the Arch document, and the 2nd is not (Dan C.) * Clarifying that we are talking fundamentally about what language specifications should provide for, and only secondarily about what individual instances should or should not contain (Henry T.) * Discussing the example of XML 1.1 (Dan C. -- FWIW, I think this is useful, but it also makes that section a bit longer and lumpier than I'd like. I may keep it or may delete again before publishing. Strong opinions?) * Added links to Dave Orchards blog postings on versioning (Thanks, Dave!) * Deleted the suggestion that when you include v1 and v2 in a list of versions your instance works with, that you're then committed to revising the instances if a compatible v3 spec comes out (Marc de Graauw) * Making clear in the signature that I'm speaking for myself as opposed to for the TAG. * Maybe other small changes I'm forgetting. I'm afraid I did not take Mark Baker's suggestion to discuss media types. They are an important aspect of versioning on the Web, but the scope of this note is the pros and cons of having language specifications provide for version information in band in the document. Media types are an example of a marker contained in a container (HTTP) and referring to the type or version of the contained data. That's important, but different I think. I believe I also received a suggestion to mention HTML DOCTYPEs, and I did not act on that because it wasn't clear that they are sufficiently different in abstract architecture to merit space in this already too-long blog entry. This is not intended to be nearly as comprehensive as a finding. So, I hope the TAG and www-tag readers find this agreeable. In principle, I suppose I can post it even if you don't, but in practice I welcome comments, and it looks like the blog won't be there for at least a bit anyway. Thanks! Noah [1] http://www.w3.org/2007/08/13-tagmem-minutes#item04 [2] http://www.w3.org/2001/tag/2007/08/versionBlog-20070806.html [3] http://www.w3.org/2001/tag/2007/08/versionBlog-20070904.html [4] http://www.w3.org/2001/tag/2007/08/versionBlog.html -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 --------------------------------------
Received on Tuesday, 4 September 2007 22:16:16 UTC