Re: A modest hyperlinking proposal

Hi Micah,

>>I take it, then, that the difference that you were referring to
>>between <a> (xml:href) and <img> (xml:src) was one of actuation
> I think of it like CSS. HTML has a "default stylesheet", which
> provides guidance on how different elements should be styled.
> Implementations are free to follow or diverge from the default
> stylesheet in whatever way makes sense.
> Similarly, one common interpretation would be:
> xml:href => xlink:actuate="onLoad" xlink:show="replace"
> xml:src  => xlink:actuate="onRequest" xlink:show="embed"
> but a major point is not to hard code it.

Oh. So in other words, the two attributes give you two possible
actuate/show combinations to play with, out of the possible 6 (or 8 if
you include actuate="onRequestSecondary"; I'm not counting 'other' or
'none'). Which combination you associate with a particular attribute
on a particular element is up to you. So for example, I could just as
easily do:

xml:src  => xlink:actuate="onLoad" xlink:show="replace"
xml:href => xlink:actuate="onRequest" xlink:show="embed"

Have I interpreted you correctly?

I thought you were going for something deeper than providing two
possible attributes rather than one -- that there was something that
actually distinguished them in terms of when each should be used. If
not, then I think that the suggestion (was it Eric's?) of turning
*any* attribute into a link by putting it in an XLink (or something)
namespace would be better. That way you could do:

  <img xlink:src="..." xlink:longdesc="..." />

and, indeed, have as many attributes as you wanted, with whatever
(local) names you wanted, specifying behaviour and semantics based on
the local name somewhere else.


Jeni Tennison

Received on Monday, 30 September 2002 17:43:30 UTC