Re: efficient XML (!= recursion)

["Michael Champion" <mike.champion@softwareag-usa.com>]

----- Original Message -----
From: "Eckenfels. Bernd" <B.Eckenfels@seeburger.de>
To: "ebXML-Architecture List" <ebXML-Architecture@lists.oasis-open.org>
Sent: Monday, March 13, 2000 3:47 AM
Subject: RE: efficient XML (!= recursion)


> > Tim Bray's criteria - that an XML parser should be easily
> > implementable with two to three weeks work by a
> > graduate student - should not be forgotten in this context.
>
> Requiring to build your own Parser (like one which will supports
> BizCodes) is too much work already. ebXML should require only SAX or DOM
> parsers.. otherwise you wont have any advantage over EDIFACT (which you
> can poarse within a few days with a hand build parser, too).

EDIFACT users need hand-built parsers, ebXML users may find them useful.

As I understand the argument here, people arguing for keeping the ebXML
schemas minimally complex are saying that it *should be possible* to write a
custom-optimized parser in a few days.  It will always be possible to use
SAX/DOM parsers off the shelf, but they may be larger, slower, or less
reliable than a particular business needs., or a particular application may
run on a legacy system for which SAX or DOM parsers are not available. The
more features of XML that ebXML uses, the harder it will be to write an
efficient parser by hand.