W3C home > Mailing lists > Public > www-html@w3.org > January 2007

Re: XHTML2.0 - transclusion

From: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Date: Wed, 24 Jan 2007 09:50:28 +0000
To: Jakub Dabrowski <jakubdab@gmail.com>
Cc: "www-html@w3.org" <www-html@w3.org>
Message-Id: <1169632228.5776.21.camel@galahad>

Jakub Dabrowski asked:
> For example ? 

Well, let's take your example (application/xhtml+xml to text/plain)
even. What specification defines how:

<a href="http://www-tech.mit.edu/Shakespeare/hamlet/hamlet.1.4.html"
title="Hamlet (MIT Online Edition), Act 1, Scene 2">the ghost scene in
Hamlet</a>

should be represented in text/plain?

> For today, until someone get better idea - yes. This is because
> someone can try to include content which tries to include including
> content ;) They'll both wait for each other and that's really stupid
> thing... 
> Eg.
> <!-- document with translusion: doc1.html -->
> <object src="www.remoteserver.com/doc2.html" id="main"
> srctype="application/html+xml" type="plain" transid="someid">
> <em>cannot get content from remote server</em>
> </object>
> 
> > <!-- other document with translusion: doc2.html -->
> > <object src="www.remoteserver.com/doc1.html" id="someid"
> > srctype="application/html+xml" type="plain" transid="main">
> > <em>cannot get content from remote server</em> 
> > </object>

User Agents will be aware of the order of transclusion: that is which
URI called which URI. They will therefore be able to tell when they
encounter a circular reference to an earlier URI in the chain.

Therefore, rather than cancelling any transclusion which includes a
sub-transclusion regardless of whether it involves a circular reference,
why not:

1) Cancel the entire transclusion when a circular reference is
encountered.

2) Cancel only the circular inclusion, loading the rest of the
transcluded content normally.

You didn't comment on my suggestion of replacing transid with #.

Thinking about "<em>cannot get content from remote server</em>",
shouldn't fallback content be a genuine fallback? Shouldn't messages
about being unable to load content be left to User Agents to construct
in a consistent fashion? "Cannot load remote image" would not usually be
good alt attribute text.

--
Benjamin Hawkes-Lewis
Received on Wednesday, 24 January 2007 09:53:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:08 GMT