RE: Formatting Objects considered harmful

Paul Prescod wrote:

> There is a subtletly that we are passing over here. When a client sends me
> information in their corporate DTD it takes me about half a day to wrap my
> head around it enough to be able to do something useful with it. You can't
> throw it into a search engine or word processor and expect to be able to
> do anything useful with it. To make it useful, either the customer or I
> must map it to some more generally known data model -- either a display
> data model like FOs, or a search model.

	This is all true, but interface building is one of the most significant
activities going on in the real world. The issue isn't whether or not
interfaces need be built (because this isn't a perfect world) rather given
your druthers, would you rather build an interface to a system which i/os
XML expressed via a DTD or, alternatively a poorly documented binary
protocol. Half a days work seems like a very short time ... Having done my
share of systems building, it seems that a common activity is transforming
on set of data onto another, gluing together libraries, matching APIs,
converting a query in one language to SQL etc etc etc. To the extent that
the I/O is via XML and an arbitrary DTD, much of this transformational
interface building can be expressed via XSLT. This has important
implications for tasks such as EDI.

	Suppose you had the desire to build an interface engine which accepts
various types of EDI messages, and suppose an XML converter exists for each
message format but each company uses a slightly different format for its
messages, how would you go about designing such a system? Would you express
the message format conversions in XSLT, perl, java, or C++?

Jonathan Borden
http://jabr.ne.mediaone.net



 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list