- From: Philip Jägenstedt <philipj@opera.com>
- Date: Thu, 04 Nov 2010 11:02:04 +0100
- To: "Media Fragment" <public-media-fragment@w3.org>
<http://www.w3.org/2008/WebVideo/Fragments/tracker/issues/19> I've drafted what I think the first steps towards resolving ISSUE-19 should look like: <http://people.opera.com/philipj/2010/11/04/media-fragments-spec/overview.html> The main changes are: * I've split the syntax into two levels: (1) the general name-value structure (namevalues production) and (2) the dimension syntaxes, operating on the Unicode strings extracted from the first pass. I've also been clear(er) on what components are, and note that as a result the URI spec requires percent-decoding to happen after parsing out the names and values. * I've reworked the section "Processing name-value components" and "Processing name-value lists" and made them normative by moving them into section 5. They still contain algorithms, but they are much smaller than before. If anyone can express the same thing in prose, that would be fine. * I've removed the production rules that are not used and were in any case nowhere in the normative part of the spec before I began editing. This includes, notably, the mediasegment production. What's left to do: * The dimension sections need to be cleaned up and be given better hooks for section "Processing name-value lists", so that it is abundantly clear what happens with e.g. #t=10&t=20. * There needs to be stronger language about validity. We should state that any unrecognized names or values are invalid, as is repeating the same dimension twice. To make it simpler to review the changes, I've also prepared a set of patches that apply to WWW/2008/WebVideo/Fragments/WD-media-fragments-spec in CVS. <http://people.opera.com/philipj/2010/11/04/0001-Fix-indendation-of-timeprefix-timeparam-syntax.patch> <http://people.opera.com/philipj/2010/11/04/0002-Change-odd-use-of-fragmenting-to-addressing.patch> <http://people.opera.com/philipj/2010/11/04/0003-Replace-uses-of-field-value-with-name-value-for-cons.patch> <http://people.opera.com/philipj/2010/11/04/0004-Fix-RFC-3986-typo-was-3896.patch> <http://people.opera.com/philipj/2010/11/04/0005-Define-parsing-normatively-and-similar-to-query-comp.patch> <http://people.opera.com/philipj/2010/11/04/0006-Clean-up-and-update-ABNF-references.patch> <http://people.opera.com/philipj/2010/11/04/0007-Replace-utf8string-with-unistring.patch> <http://people.opera.com/philipj/2010/11/04/0008-Purge-some-unused-production-rules.patch> 0005, 0007 and 0008 are the most important, the others are fixing small issues. Please take the time to review this so that we can perhaps finally move on with ISSUE-19. -- Philip Jägenstedt Core Developer Opera Software
Received on Thursday, 4 November 2010 10:01:18 UTC