W3C home > Mailing lists > Public > public-html@w3.org > July 2008

RE: On ISSUE-41: Decentralized extensibility

From: Chris Wilson <Chris.Wilson@microsoft.com>
Date: Mon, 21 Jul 2008 10:39:51 -0700
To: Thomas Broyer <t.broyer@gmail.com>
CC: "public-html@w3.org" <public-html@w3.org>
Message-ID: <D12127075745E648BBC075EF46983E17102BCDF74E@TK5-EXMBX-W603v.wingroup.windeploy.ntdev.microsoft.com>
Thomas Broyer [mailto:t.broyer@gmail.com] wrote:
><Chris.Wilson@microsoft.com> wrote:
>> http://msdn.microsoft.com/en-us/library/ms531426.aspx#Creating_Custom_Element is the best reference I could find.
>
>Sigh. Being a "user" doc, it doesn't deal with how mismatched tags are
>handled (you're supposed to match your tags after all)

Nope, we didn't really write developer docs on this stuff.

>Would it mean you do not have any internal doc that could be published
>as a white-paper of some sort?

Not that I can find, no.

>> From memory - /> never signals an empty element to us - we always, I think, required a close tag.
>
>That's not what the above linked document says ;-)
>http://msdn.microsoft.com/en-us/library/ms531426.aspx#ParsingEBTags


Hmm, right you are.  I guess I just tried to block that whole feature out of my mind.  :)

>>  And we didn't really check "wellformed" very well.  :(
>
>Of course, and that's kind of a feature IMO (remember XML5? "XML with
>error recovery"?); my question was "what does IE do with
>"non-wellformed" content?".
>See the last 4 lines of http://www.ltgt.net/ie-namespace-tests.html

>for "empirical reverse engineering".

Ah, I see.  I'd not focus too much on how IE6-7 works in this regard, and focus on how we WANT it to work.

There are obvious missing tests here, re how HTML tags like P or
TABLE, or "formatting elements" behave wrt "custom elements": do <p>
or <table> imply end tags for custom elements? how about </p> ? and
unmatched <b> or <i> in the middle of unmatched custom elements? etc.
I'll try to find time to update the page with such tests (and try to
fix the page so that "view source" shows the actual source, not the
generated one)

--
Thomas Broyer

Received on Monday, 21 July 2008 17:40:56 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:56 UTC