RE: typing

[Jonathan Borden <jborden@mediaone.net>]

I wrote:
>
> The surprising number of perhaps subtle but real errors I've
> found in common
> XML parsers related to validation leads me to strongly suspect
> that not much
> validation is going on in common everyday practice. I'm not sure
> how this is
> a real problem in terms of complexity.

I shouldn't have let this slip without qualification.

An example: try to browse:

http://www.openhealth.org/ASTM/operative.report.example.xml

in IE5 (I'm using version 5.50.4522.1800):

It hangs!

This is due to external parameter entities at URLs within the DTD which is
at a URL (i.e. the same document validates and browses just fine when the
document and DTD and entities are on the local disc).

Seems that not alot of work has gone into testing the IE5 DTD related code
under a large variety of conditions -- perhaps this reflects the apparent
fact that not alot of Microsoft's customers think this is important.

(note that the offending modules are XHTML Basic -- so don't blame us in the
ASTM XML Healthcare committee -- except perhaps for wanting to use XHTML
Basic!)

Other validators such as the W3C's fail to process URL redirections etc.
etc. (as I said subtle issues that might -only- come out if the systems were
being extensively used). Just my opinion.

Jonathan Borden
The Open Healthcare Group
http://www.openhealth.org

and Co-Chair ASTM E31.25 XML Healthcare DTD Standards
see http://www.openhealth.org/ASTM/, http://www.astm.org