Re: Argh...Entities

At 09:22 AM 5/18/99 -0500, Paul Prescod wrote:
>MURATA Makoto wrote:
>> 
>> On the other hand, we cannot use SAX and DOM as is.  
>
>As is? No. But we can build on them. We have to build on them to make
>schema parsers anyhow.

We _can_ build on them - but I have concerns that the (significantly larger
and more complex) schema processing is going to get lumped in with the
current XML 1.0 mechanism and give us bloated parsers that are difficult to
control.

I'd love to see the schema spec people describe in some fairly scandalous
detail how exactly schemas should interact with XML 1.0 documents - and
make sure that all schema-compliant docs are 1.0-compliant.  I've made some
proposals in my Layered Model article
(http://www.simonstl.com/articles/layering/layered.htm) and XML Processing
Description Language (XPDL - http://purl.oclc.org/NET/xpdl ).  I'd like to
see the schema WG and supporting groups make the connections between the
various layers.

>And if a particular application does not need the extended schema
>information then it can work directly with the SAX 1 or DOM level 1.

So we hope.

>> Existing XML infrastructure will not work any more.
>
>We're just talking about a layer on top. XLink doesn't invalidate XML 1.0.
>It just means you need another layer of tools to get the full meaning of a
>document.

As noted above, however, we need to make sure that the layering of those
tools is well understood and that layers do not interfere with each other.
I don't think we're nearly there yet.

Simon St.Laurent
XML: A Primer / Building XML Applications (June)
Sharing Bandwidth / Cookies
http://www.simonstl.com

Received on Tuesday, 18 May 1999 12:18:56 UTC