RE: Forest Automata formalism

Reynolds, Gregg (greynolds@datalogics.com)
Wed, 21 Oct 1998 09:40:50 -0500


Thanks to Paul for making the paper available. I've been working on a
tree query language that has ended up looking much like a forest/tree
regexp language; I had not considered it from the dtd/schema
perspective. Looks very promising!

FYI: There is a bibliography of 144 items in "Handbook of Theoretical
Computer Science Part B", Elsevier and MIT Press 1996, Chapter 4
"Automata on Infinite Objects" by Wolfgang Thomas. Part Two of the
chapter is entitled "Automata on Infinite Trees." Caveat: very heavy
sledding. Most of it sails way over my head, but on pages 165-166 there
is a very concise account of tree concatenation etc., which I found
understandable and useful. You can probably find this in a bookstore if
you're in a major city.

Two other good references:

"Text Algorithms" by Maxime Crochemore and Wojciech Rytter, Oxford 1994

"Algorithms on Strings, Trees, and Sequences: Computer Science and
Computational Biology", by Dan Gusfield, Cambridge U. Press 1997. Very
cool book. Excellent and thorough treatment of suffix trees.

- gregg

> -----Original Message-----
> From: Paul Prescod [SMTP:papresco@technologist.com]
> Sent: Tuesday, October 20, 1998 10:50 AM
> To: xml-dev; tigue@oz.net; macherius@darmstadt.gmd.de; Sean Mc Grath
> Subject: Forest Automata formalism
>
> Several people have asked me about my paper on Forest Automata.
> Unfortunately, I haven't been able to find time to compile a decent
> bibliography, nor add the code examples I hoped to, but I'm going to
> release it anyhow. There is a partial(?) bibliography on Robin Cover's
> SGML/XML page.
>
> My paper (sans biblio) can be found at:
>
> http://www.prescod.net/forest/shorttut
>
> Forest Automata theory is a formalism for describing SGML/XML
> validation.
> The formalism makes clear some obvious extensions to SGML/XML
> validationand can be used as a source of ideas and answers relating to
> DTD
> parameterization, "data types" (lexical data types),
> validation-in-context, query languages etc.
>
> Hopefully I will find time or funding to expand this paper eventually,
> because there are many interesting highways and biways that it hints
> at
> but does not explore.
>
> Paul Prescod - http://itrc.uwaterloo.ca/~papresco
>