OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   Re: interoperability (was Re: Obfuscating XML with namespaces)

[ Lists Home | Date Index | Thread Index ]
  • From: David Megginson <david@megginson.com>
  • To: xml-dev@lists.xml.org
  • Date: Mon, 09 Oct 2000 17:24:27 -0400 (EDT)

Dave Winer writes:

 > My pov: XML should be providing a means for interop between all kinds of
 > apps. It should be happening now.

Data exchange is still missing too many layers.  

(Everyone who's heard this before can tune out now...)

XML itself is like IP -- it's a good foundation, but it's too raw to
give us interoperability.  We need to build the equivalents of TCP,
UDP, and then HTTP, FTP, SMTP, etc. on top of it to accomplish much
more than we're managing now.

RDF (which Dave doesn't like, and I'm still wishy-washy about) is an
attempt to build a data layer on top of XML.  If we can all agree on a
data layer -- ANY data layer -- then we can build an awful lot of
reusable software for it (forms support, database and spreadsheet
import/export, browsers, search and indexing and query engines,
language-specific data binding, etc. etc.) and then build
domain-specific data-interchange specs as thin layers on top rather
than monolithic monstrosities we're vomiting out now onto an
unsuspecting public.

As it stands right now, nearly every spec that exchanges tabular,
fielded, or object-based information reinvents the wheel starting with
low-level XML, and forces people into contortions like writing XSLT
stylesheets for simple data exchange.  

RDF has way too many ugly parts, but there are some specs being built
on top of it (RSS 1.0, PRISM, XMLNews-Meta, the now-moribund INDECS,
and a few others that have been mentioned on XMLHack), so it's a tiny
spark.  We can either try to make RDF work, or throw it away and
invent something similar (but maybe without all of RDF's reification
junk and messy syntactic variation).  

SOAP and XMI are other candidates for a data layer, but while SOAP
makes good use of Namespaces, it mixes up procedural RPC too much with
data representation (read: BIG LAYERING VIOLATION, as if HTTP and HTML
were defined in the same spec), and XMI is so Byzantine that it makes
the RDF spec look read like The House at Pooh Corner.  We need to keep
thinking.


All the best,


David

-- 
David Megginson                 david@megginson.com
           http://www.megginson.com/




 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS