• From: rjelliffe@a...
• To: "Michael Ludwig" <mlu@a...>
• Date: Tue, 22 Sep 2009 01:49:17 +1000 (EST)

> Isn't there a desire in XDM to get away from the nitty-gritty XML syntax
> details and join the realm of programming languages and their data
> structures living in memory?

This sounds like you think some other serialization won't have syntactical
nitty-gritty!

If your idea is that the XPath Data Model is an improvement on XML, then
you are lost in the realm of apples and oranges, no slight intended.

> In all cases, will it not be more efficient to just use some
> run-of-the-mill serialization format? I'm sure plenty of those
> have been produced to foolproof perfection throughout the decades.
> I would honor those efforts instead of rolling my own wheel in
> angle brackets.

Hey, maybe someone should suggest JSON, as the only
programming-language-independent data-oriented serialization format with
traction!

I note that Charles Goldfarb had an inlining model for XML type
information before XSD, in maybe 1997: IIRC it would be (updated with
namespace) something like

 <address city="NY" type:info=". ex:addressType city ex:cityList">
    ...
where there is an added attribute with name/type-name pairs.

And note that ISO NVDL handles disconnected attributes by making a special
dummy element whose semantic is strictly to be a dummy element.

Cheers
Rick Jelliffe

• Follow-Ups:
  • Re: Serialization of XDM
    • From: Michael Ludwig <mlu@a...>

• References:
  • Serialization of XDM
    • From: "David A. Lee" <dlee@c...>
  • Re: Serialization of XDM
    • From: Michael Ludwig <mlu@a...>