Re: Some questions on HLink

From: "Tim Bray" <tbray@textuality.com>

> I read HLink in detail and had the following questions; quite likely
> others will as well so I might as well post them here.
>
> - I note the Status of this Doc says "...does not imply endorsement ...
> of members of the HTML Working Group."  Since lots of W3C WDs do not
> carry this disclaimer, I wonder if it is just boilerplate or not.

It is indeed boilerplate. I am not aware of any non-endorsement within the
HTML (or Forms, who are also tracking it) working groups.

> Is this a trial balloon, or something that the WG is pretty well agreed
on?

It is something which we have been discussing for quite a while - we are
pretty well agreed on it; we are also in communication with the CSS group
for additional properties for styling links.

> is the default action that this shows up in XHTML 2.x?

I don't quite understand this question. Which default action do you mean? We
regard Hlink like CSS: we would describe our links using it, but not require
a UA to implement it.

> I note there are open issues; does consensus exist outside of these?

I believe so.

> - Just a quibble: I think what the WD labels as "implementations" would
> better be described as "declarative schemas" or "formal descriptions" or
> some such.  Or am I missing something?

This is 'Modularisation of XHTML' talk.
(http://www.w3.org/TR/xhtml-modularization/)
This talks about 'abstract modules', and then implementations of those
abstractions in different schema languages.

> - Is there any write-up anywhere of how this differs from the ISO
> Architectural Forms framework?

No, I'm afraid not.

> - There's a formatting problem: I assume that the possible values of
> "effect" are new, replace, embed, submit, and map?

Hmm, it looks alright on my screen. Which browser are you using? Yes, those
are the values.

> - The definition of the "replacement" attribute uses the undefined term
> "environment"; the (only) example in 2.3.1 shows that this is, in the
> HTML context, a frame.  I'm trying to imagine portable non-HTML uses of
> this attribute.  Are there any in mind?

Well, it is trying to be as abstract as possible: whatever works in the
environment that the link is actuated in. In the HTML world that is indeed
Frames and Windowing systems, but we don't want to restrict the use of the
'replacement' to only those, but any that works.

> - It might be the case that in some circumstances you might want to
> design a multi-ended link using subelements rather than attributes to
> carry the URIs and so on (as in the XLink extended links).  HLink (if I
> read it right) limits you to using attributes for this.  Is there any
> thought to extending this to element-based designs?

It's not out of the question. We are chartered to solve the problem for
XHTML, so we therefore limited the design for our needs. I would be happy
for extensions to be added if that made it more widely applicable.

> - On section 3, would you envision that in some subsequent version of
> XHTML, the "a href=" element would have no linking effect unless the
> hlinks are explicitly included in or referenced from the document?

No; we would envision a usage like that of CSS.

Best wishes,

Steven Pemberton

Received on Wednesday, 18 September 2002 05:43:57 UTC