Re: Subset Data Model

On Mon, Aug 13, 2012 at 4:56 PM, David Carlisle <davidc@nag.co.uk> wrote:

> On 13/08/2012 23:25, Murray Maloney wrote:
>
>> My existing XML 1.0 tools will be able to read uXML files. A future
>> generation of uXML tools may hiccup on raw XML 1.0, especially if
>> resorts to DTDs and entities.
>>
>
> Yes but David's (valid) point is that is a _bad_ thing if the only thing
> you are guaranteed is the syntax.
>
> If
>
> <a b="x"/>
>
> is well formed xml and micro-xml but
>
> if parsed with xml parser generates an empty element with attribute b
> with value x
>
> and
>
> if parsed with a micro-xml parser produces the element x with text content
> b
>
> then the fact that I can develop a micro-xml application and someone
> with only an xml parser can parse the file is not clearly a good thing.
> It might be better to use a different syntax and get a fatal error if
> you parse it with an xml parser.
>
> Of course the xml 1.0 spec on its own says nothing much about the data
> model and this is on the whole a good thing, but you need _something_ at
> the level of data-model (or infoset) (or something) if you want to say
> micro-xml has any relationship at all to XML.


David, I don't think reductio ad absurdum works here, as John indicated
before.

To follow the example you suggest, XML 1.0 did not even specify the order
in which elements should be reported by a parser.   In theory the XSLT 1.0
model could have said that node sets from a node match appear in
alphabetical order of string value, rather than document order.  This would
not have violated the syntax of XML 1.0, but it would have been absurd. It
would have made it nearly impossible to do anything useful in that
processing.

So the XSLT/XPath 1.0 data model did what was sensible and useful (for the
most part).  The MicroXML data model shall do the same.  I have heard
nothing whatsoever to suggest otherwise, which is precisely why this
reductio is merely tedious.

What we *have* said is that we will not define what is sensible and useful
by strict derivation from any present data model spec.  Of course a
difference in behavior between MicroXML and Infoset, and heck, maybe even
XDM could count as an argument against any particular choice. Our point is
simply that it's an argument that has to be made.  That will be
facilitated, once again, once we actually have a straw man MicroXML data
model to throw darts at.


-- 
Uche Ogbuji                       http://uche.ogbuji.net
Founding Partner, Zepheira        http://zepheira.com
http://wearekin.org
http://www.thenervousbreakdown.com/author/uogbuji/
http://copia.ogbuji.net
http://www.linkedin.com/in/ucheogbuji
http://twitter.com/uogbuji