[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Designing XML to Support Information Evolution

• To: "'Rick Marshall'" <rjm@z...>,"'Michael Champion'" <mc@x...>
• Subject: RE: Designing XML to Support Information Evolution
• From: "Michael Kay" <michael.h.kay@n...>
• Date: Fri, 21 May 2004 09:08:43 +0100
• Cc: "'xml-dev DEV''" <xml-dev@l...>
• In-reply-to: <40AD44D1.4060509@z...>
• Thread-index: AcQ+xZOoeeNPBxxpSWSXCZXFhAz8mgAQdQEg

> hierarchies fail, and this is my struggle with xml at the 
> moment, when they have to support multiple hierarchies simultaneously. 

Perhaps it's time to rediscover the network (or Codasyl) data model, which
was invented to solve this problem. I have always felt this model is the one
that comes closest to a direct representation of the structures that come
out of data modelling. In XPath terms, it just means that instead of having
one child axis and one parent axis, an element can have any number of
different, named, child and parent axes. It's easy enough to represent in
XML, you just use ID and IDREFS attributes for the "non-primary"
hierarchies.

The problem in this area is that XML has too many different ways to
represent relationships, and none of them is especially well supported by
declarative mechanisms in the schema or query languages. 

Michael Kay

• Follow-Ups:
  • Re: Designing XML to Support Information Evolution
    • From: "Ken North" <kennorth@s...>

• References:
  • Re: Designing XML to Support Information Evolution
    • From: Rick Marshall <rjm@z...>

• Prev by Date: RE: Designing XML to Support Information Evolution
• Next by Date: RE: Designing XML to Support Information Evolution
• Previous by thread: Re: Designing XML to Support Information Evolution
• Next by thread: Re: Designing XML to Support Information Evolution
• Index(es):
  • Date
  • Thread