W3C home > Mailing lists > Public > www-tag@w3.org > September 2002

Re: Announcing HLink: rationale

From: Steven Pemberton <steven.pemberton@cwi.nl>
Date: Fri, 13 Sep 2002 22:18:03 +0200
Message-ID: <011b01c25b62$aa84d5b0$e73b000a@srx41p>
To: <www-tag@w3.org>, "Paul Grosso" <pgrosso@arbortext.com>

From: "Paul Grosso" <pgrosso@arbortext.com>

> The "embed" value for xlink:show has no "inclusion" semantics.
> The embedding is just at a presentation level.  It was meant
> to handle such things as embedding a graphic image.  (I have
> long argued that the "embed" value was a mistake that should
> never have been allowed in XLink.)  So Steven's examples which
> (I think) are meant to imply some kind of inclusion semantics
> (e.g., act as though the referenced script code were inline)
> could not use XLink's embed.  XInclude maybe, but not embed.

I agree that embed is vague, and this was one of the HTML WG's recurrent
comments to the Linking WG. However, when I asked the Linking WG directly
(at a Linking FtF) I got the answer from them that 'embed' included such
things as script and frames:

http://www.w3.org/XML/Group/1999/09/xll-19990927-min.htm

"StevenP: the script tag has only one locator and it points to some script;
does XLink support that?

We said that the semantics of show="embed", being media-type specific, would
support this. Likewise, frameset could also be covered by "embed".
"

On the other hand, if there is a general feeling that 'embed' is a bad
choice for scripting, then we can simply add new values to HLink to cover
such cases.

Best wishes,

Steven Pemberton
Received on Friday, 13 September 2002 16:18:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:11 GMT