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


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] The privilege of XML parsing - Data types,binary XML and

[ Lists Home | Date Index | Thread Index ]

[Roger Costello]
 >I think that I see what you are saying Sean, but I'd like to analyze this a
 >bit.  Let's take the aircraft example:

 ><?xml version="1.0" encoding="UTF-8"?>
 >    <elevation>12000</elevation>

 >There is no associated data model with this.  It is just a string.  It is 
 >form that gets transmitted from client to application?

Yes but then the application puts it through an XML processor. This results
in a data model (sequeence of events, a tree, series of tokens, whatever).

So the question is - what is in that data model? I advocate that is should
be as little as possible above the surface syntax and certainly should
not magically result in elevation becoming a 32 bit integer from
the point of view of my application.

 >I have some questions about something that you said: "Applications come 
and go
 >but data lives forever".  I believe that when you say "data" you are 
 >just to the string:

 ><?xml version="1.0" encoding="UTF-8"?>
 >    <elevation>12000</elevation>

Yes. My business has a sales ledger. Most businesses do. What is a sales
ledger? In data terms the details haven't changed much in centuries. Yet
every few years we chuck it all out and replace it with a new system - an 
Typically when we do that existing data becomes legacy overnight.

Its total madness and XML, properly applied, can help stop this madness. A key
part of doing it is to make the data as independent of the foibles of any 
application as we can. After all, we are going to chunk away the application
if a few years :-( I think cleanly separating the fundamental data  from
any particular applications interpretation of it, is an important survival

Again, as I said before, I'm not against data typing but I see it as being
an application level - not a parser level view of the data. I think
pipelines can neatly address the issue of how and where datatyping
gets added so that when an application goes up in smoke but
its data is retained, we can easily get at the data as distinct from
some applications view of that data.




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

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