- From: Sam Ruby <rubys@intertwingly.net>
- Date: Thu, 27 Aug 2009 05:42:54 -0400
- To: Robin Berjon <robin@berjon.com>
- CC: Anne van Kesteren <annevk@opera.com>, Maciej Stachowiak <mjs@apple.com>, Doug Schepers <schepers@w3.org>, public-canvas-api@w3.org, "public-html@w3.org WG" <public-html@w3.org>
Robin Berjon wrote: > On Aug 18, 2009, at 08:55 , Anne van Kesteren wrote: >> On Tue, 18 Aug 2009 08:49:57 +0200, Maciej Stachowiak <mjs@apple.com> >> wrote: >>> The parts of HTML5 defining things like the Window object and origins >>> could in theory be factored out, but it's decidedly nontrivial to do so. >> >> I think it would be awesome if someone would take the effort to split >> Window, origin, storage mutex, event loop, navigation, page loading, >> etc. out. > > Admittedly I haven't looked closely enough, but I thought that a bunch > of existing documents were generated from the same source. Assuming that > would be the right way to do it, what sort of effort would be involved > in producing another generated document? Take a look at: http://svn.whatwg.org/webapps/source Specifically, look for "<!--START " and "<!--END". The overall process is pretty simple. Simple script selects the specific sections based on the presence of these markers. A header is slapped on the front containing the title page, etc. The result is first passed through Anolis, and then (optionally) spec-splitter tools. Running the full process on the full HTML5 spec takes me less than a minute on inexpensive commodity hardware. If you or anyone else can focus on what the content should look like, there are plenty of people who can help with the mechanics. So... if you or anyone else will do the "awesome" part, I'll help with the mundane part. For those interested in the technical details, the script I use is here: <http://intertwingly.net/tmp/html5spec>. The part that selects the specific document is the four lines that start "perl -e". - Sam Ruby
Received on Thursday, 27 August 2009 09:43:39 UTC