RE: fragment exchange (was Re: rationales for TEI extended-pointe r keywords)

On Sun, 15 Jun 1997 00:31:17 -0400 (EDT) Jon Bosak said:
> ...   It is very easy to visualize an
>application like JUMBO that is doing useful things with XML and needs
>links but is not doing anything that would require a style language.
>Saying that an application has to support a subset of DSSSL just to be
>able to use links is not acceptable, and we have already concluded
>that to use links minimally you have to have the SHOW and ACTUATE
>settings.  So you will have those two attributes in xml-link no matter
>what you get later in xml-style.

It should probably also be pointed out that a number of participants
argued strongly that information providers need the ability to
associate certain fundamental intentions with regard to their links.
This applied in particular to links that should always, in normal
browsing, be traversed without user intervention and displayed to
the user.  True, a style sheet in effect for a particular
application (e.g. a robot web crawler) might not always obey such
owner intentions; this was recognized and held irrelevant to the
key issue, which was the information provider's ability to give
an indication of the intentions.  At this level, these attributes
*do* inhere in the link rather than in the style, and so are
appropriately defined as attributes of the XML architectural forms.

-C. M. Sperberg-McQueen

Received on Monday, 16 June 1997 14:57:52 UTC