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] Automated schema migration

[ Lists Home | Date Index | Thread Index ]

Hmm.  Is there is an optimum trading strategy 
for schemas among trading partners with dissimilar or 
similar schemas where the payoff is minimum transformation 
of one or another document collection?  In other words, 
given multiple partners who have started independently 
and now want to merge their collections, what is the 
optimum strategy for trades of production for production? 

Will degrees of similarity change the strategy?

Are some productions of higher value than others 
(similar to playing poker)? 

If treated as a game, how many Nash equilibrium exist?

What kind of game is this?  (given perfect knowledge 
vs. incrementally acquired knowledge, competition vs. 
coordination:  remember, standards are a coordination 
game to create a condition of a stable equilibrium)

Given closed and open formats, this may not be a  
merely theoretical thought experiment.

Game theorists?  Schema negotiators?


From: Michael Kay [mailto:mike@saxonica.com]

> My question is: assuming I've created a schema for my old set 
> of documents, 
> if I were to create a transformation T1 that, given the old schema, 
> generates the new schema, could I machine-generate from that 
> transformation 
> T1 another transformation T2 that could be applied to the documents 
> themselves?

It's easy to demonstrate that this is impossible in general. Given a schema

<p>Some <i>text</i> in italics</p>

and another schema for

<para>Some <italic>text</italic> in italics</para>

It's clearly impossible to deduce the transformation by looking at the
schemas alone.

There are plenty of tools out there (generally called "mappers") that allow
you to design a transformation starting from two schemas, by showing the
relationships between the elements in each. They only work where the two
schemas are quite similar, but that's probably true in your case.

Michael Kay


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

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