• To: "'Joe English'" <jenglish@f...>, xml-dev@l...
• Subject: RE: (data) medium is the message
• From: Jeff Lowery <Jeff.Lowery@c...>
• Date: Thu, 1 May 2003 09:52:43 -0700

> Maybe not completely compatible, but certainly complementary.
> Google for "XML normal form," a lot of people are working
> on this.

I'll be danged if I didn't independently write about this same topic well
over a year ago.  Unfortunately, it was in chapter of a book that appears
fated never to see the light of day...  <sigh>

I wasn't as formal as
http://www.cs.ualberta.ca/~yuan/courses/692/lectures/xnf/xnf.pdf, either.
But it was basically the same method of applying functional dependency
analysis (the motivation for relational normalization) to XML documents.

The thing is, though, is that you don't want all your XML documents to be
normalized.  The point of using a mapping is that you can note where an XML
source is not normalized (e.g., components that contain replicated data),
and recognize it as such when porting data over to a relational model.  And
you'll want to check that the XML data was truly replicated faithfully.

• Follow-Ups:
  • Re: (data) medium is the message
    • From: "Bob Foster" <bob@x...>

• Prev by Date: rfc 1738
• Next by Date: RE: (data) medium is the message
• Previous by thread: Re: (data) medium is the message
• Next by thread: Re: (data) medium is the message
• Index(es):
  • Date
  • Thread