[
Lists Home |
Date Index |
Thread Index
]
Simon St.Laurent wrote:
> You may find that the Infoset isn't as popular a notion as you seem to
> find it, and that the only Infoset extensions some people (myself
> vocally included) are willing to work with are those which can be
> expressed explicitly in XML 1.0.
i honestly don't see the point. people working with dom or sax or xslt
are working with the infoset every day, even though they may not know it
or call it this way. and of course it is no problem to map any infoset
extension to xml syntax, if you want to. the fact that some people don't
bother to do it (like xml schema's psvi) does not say anything about the
quality or content of the infoset extensions.
> The proposal in XML Linking and Style [1] defines XLink contributions to
> the Infoset, but it does so in terms of Element Information Items,
> adding nothing fundamentally new to the Infoset. I'm not sure whether
> your proposal and that proposal are similar.
they are similar in spirit, but the proposal you are referencing has
some shortcomings which make it a subset rather that a 1:1 mapping of
the xlink data model. and the fact that they are using "element items"
for the link information in my view is simply wrong, because element
items are for elements...
> That proposal seems fine to me because of its grounding in XML. I won't
> touch an Infoset-only proposal with a thousand-foot pole at this point;
> in fact I'd much prefer to see the XML Linking and Style proposal
> defined as a transformation that produces one flavor of XML from another
> rather than wasting cycles on Information Items.
why is that proposal "grounding in xml"? show me how to access the link
information with vanilla xpath? or using css selectors? or via dom? it
will not work, because they are *extending the infoset* and this is the
right way to go. in my view, their re-use of "element items" simply was
an attempt to make people feel more comfortable with the whole approach
(and it obviously worked ;-), but if you really want to support infoset
extensions, then you have to go all the way.
cheers,
erik wilde - tel:+41-1-6325132 - fax:+41-1-6321035
mailto:net.dret@dret.net - http://dret.net/
computer engineering and networks laboratory
swiss federal institute of technology (eth)
* try not. do, or do not. there is no try. *
|