[XML-SIG] parsers and XML, note re: xmlpickle

travish travish@realtime.net
Thu, 10 Aug 2000 19:22:38 -0500 (CDT)

> A processor that checks for correct syntax and stops there is still a
> "parser", as is a tree-builder

Validating parsers definitely qualify as "real" parsers.
DOM stuff definitely qualifies as a parser.

I guess I should have singled out "event-driven XML parsers" as
a misnomer.  I didn't mean to disparage the others.

> or a event-stream generator.

Event-stream generators emit events which loosely map to tokens.
That is, "<HEAD>foo</HEAD>" comes out as start_tag('HEAD'),
character_data('foo'), end_tag('HEAD').  They don't appear to do anything
related to the structure of the HTML.  A lexer would generate
a token stream such as TOK_BEG(HEAD) STR(foo) TOK_END(HEAD).
I'm still awaiting an explanation of how these differ in any significant
way, but I'm tired of this thread and am not going to clarify again.


Incidentally, the "nested data structure" comments are imminently
relevant to XML pickling.  Do you really want to re-code a parser stack,
reduction dispatch routines and child tree access code for every such
application?  Nor would you necessarily want to read it all and generate
an AST in memory.
Those who will not reason, are bigots, those who cannot,
    are fools, and those who dare not, are slaves.
       - George Gordon Noel Byron (1788-1824)