Lists Home |
Date Index |
From: "Dare Obasanjo" <firstname.lastname@example.org>
> A couple of people I've met (most of whom happen to work with or for the web services team at MS) have described the infoset as what is important in XML since it describes the data model while relegating the syntax described in the XML 1.0 REC  as one potential serialization of this data model (aka the XML information set). Thus it is possible to create alternate syntaxes for XML 1.0 such as Comma Seperated Value files or S-expressions which satisfy the XML infoset as long as some mapping is provided.
But CSV, S-Expressions and text/rich were around for a long time,
without getting the same kinds of pan-domain adoption that XML
is getting. Part of the reason is some features (adequate i18n for
a start) could indeed be replicated. But a lot of the benefit is simply
because paying attention to syntax/superficial structure is
a basic discipline which catches many of the real problems that
The basic XML infoset is little more than later LISP's (cons=node,
property=attribute, string=content): sure it is good, and disconnecting
it from list processing has made it available more generally, but
the infoset is hardly the whole story. I saw a great implementation of an XML parser in LISP from Japan which took only a couple of pages of code.