• From: Arnaud Sahuguet <sahuguet@g...>
• To: xml-dev@l...
• Date: Mon, 25 Jun 2001 22:01:24 -0400

Hi,

I have been using Graph for "DTD mining".
See http://db.cis.upenn.edu/DL/dtds.pdf for the paper.

I have been using GraphViz and I do not recall major problems (except time) 
even for large DTDs.
The trick is sometimes to tweak the parameters. You often do not need to 
impose ordering among nodes. It makes the placement of nodes easier.

Also, you do not always need the finest granularity. You can ignore 
attributes or put them together, depending on what you are looking for.

As for documents, I have not really played with them. But once again, 
depending on what you need, you can decide to ignore sub-trees and just draw 
them as a single node.

Arnaud

• Prev by Date: RE: "Current" list of XML IDEs?
• Next by Date: RE: XML Blueberry
• Previous by thread: RE: xml DOM/DTD graph visualization
• Next by thread: Re: Should finalDefault allow list and union?
• Index(es):
  • Date
  • Thread