Lists Home |
Date Index |
> -----Original Message-----
> From: Bill de hÓra [mailto:firstname.lastname@example.org]
> Sent: Saturday, February 22, 2003 12:58 PM
> To: Cavnar-Johnson, John
> Cc: email@example.com
> Subject: Re: [xml-dev] The subsetting has begun
> Cavnar-Johnson, John wrote:
> > This specification
> > defines a particular application of XML and the style of document
> > that application uses.
> John, you're the third knowledgeable person I've seen claim this in
> as many weeks. This is not an application of XML.
I agree. As I stated previously, my initial skimming of this document
was faulty. If the spec simply defined an API for handling web service
invocations, I wouldn't have a problem with it. But, as several folks
have pointed out, it also tries to define a non-standard approach to
parsing XML documents.
> MathML or RSS is an application of XML. And even if you and I
> couldn't agree on what an XML application is, and we chose your
> definition it still wouldn't be an application of XML. It would be
> an application of an XML Infoset. XML 1.0 is just one possible
> serialization of a SOAP message.
I'm not sure what your point is here. Are you making a statement about
this spec in particular or a larger point about SOAP?
> > As long as it doesn't purport to define a
> > generalized XML parser, why is it wrong? This application won't
> > all well-formed XML documents, but so what?
> You might want to read the document. It talks about XML and sings
> its praises, much more that it does the actual subset it defines.
> From a specification viewpoint, what is wanted is the assertion,
> upfront, that this JSR uses a subset of XML, names that subset and
> releases any claims to XML. That belongs in 1.2, Main Goals and
> Deliverables. That way, we can evaluate the JSR on its own merits,
> not XML's. Anything else is bait and switch.
> > Must every application that
> > uses XML accept any XML document?
> I think we can dismiss the argument from application.
What? I don't understand your response.
> > What's the danger here?
> Briefly it's an economic danger, not a technical one. A technical
> danger one wouldn't so much - we've known since SQL was standardized
> that developers love a challenge. But subsetting ups the cost of
> bridging systems claiming to use XML and that does matter.
After reading the spec more closely, I think the real danger is in
mandating parser behavior that will guarantee incompatibility with other