> P.S. How do browser authors feel about handling full parse trees?
> What's the average performance penalty for actually parsing the
> document using a true SGML parser? Any ideas on how to handle
> parsing before a document is completely loaded? (Should heuristics
> based on existing HTML practice be used, or is it simply too
> unknowable?)
Actually, there should be very little performance hit, if any -- an
SGML parser does not have to validate the document which it is
parsing, and building a parse tree of what the browser thinks is the
document's structure should be trivial.
It should also be possible to display the document as the parse tree
is being constructed, unless something in the stylesheet (whatever its
format) requires a forward reference. As soon as an element begins
you already know all of its ancestors, how deeply it is nested, and
all of its left siblings. If a forward reference is required (ie. you
are looking for the second-last <h2> in the document), then the parser
can build the tree up ahead only as far as necessary, or it could go
back and adjust the styles after the parse is complete.
David
-- David Megginson Department of English, University of Ottawa, dmeggins@aix1.uottawa.ca Ottawa, Ontario, CANADA K1N 6N5 ak117@freenet.carleton.ca Phone: (613) 562-5800 ext.1203 WWW: http://www.uottawa.ca/~dmeggins FAX: (613) 562-5990