RE: !BEHAVIOR

Firstly, whatever insight I have comes from a long, thorough and very 
helpful article writeen by EliotK on xml-dev.  [Suggestion - Eliot has to
explain the same thing so often, that perhaps his collected writings could
be captured and distilled.  Seriously, it's worth having a Link (SIMPLE) to.]

In message <317CDDD87D9CD011958100609712EB6B055D60@FLPS-NTSERVER1> "Rivers-Moore, Daniel" writes:
> <peter-murray-rust>
> 4.1 SHOW/EMBED suggests that a tree from a remote document can be
> embedded as part of the document in which the XML-LINK appears.  With
> ACTUATE="AUTO" this is not a stylesheet mechanism, but a way of
> combining components of documents. I have got excited about this, and am
> building this feature into CML as it is a useful way of combining
> fragments from more than one DTD.  Maybe I'm being naughty in doing
> this, but I wouldn't like to be deprived of it.
> </peter-murray-rust>
> 
> In no way do I want to deprive you of this. On the contrary, I want to
> enable this to be done in more powerful, more flexible ways.
> 
> The fundamental point is one of _indirection_. If the instruction to
> SHOW or ACTUATE in this or that way is an attribute of the element, you
> have tied it directly to the link. But the rule as to when, where and
> how the linked content should be displayed is not_a _property_

The point which I got from Eliot's article was that one use of link
technology is to regard them as structural.  IOW they complement the use of
hierarchical containment (via Element content) to provide an arbitrarily
complex document structure, and where the difference bewteen links and 
containment is syntactic only.  I wish to use these links as structural, rather
than for presentation.  It is possible that I can and should invent some ROLE
that describes this, but the use of AUTO/EMBED seemed to be generic and 
valuable.  [It is isomorphic with the IMG tag in HTML:
<IMG SRC="foo.gif"> automatically embeds foo.gif in the document at that
point.  *How* it embeds it will probably be dependent on the way you represent
your document in memory and could either be a link to a blob, or could expand
the object as part of the grove.]

[PeterMR, virtual student, HyTime 101]

-- 
Peter Murray-Rust, domestic net connection
Virtual School of Molecular Sciences
http://www.vsms.nottingham.ac.uk/

Received on Saturday, 14 June 1997 06:23:02 UTC