Re: asn1xml ASN and XML workspace

At 09:01 PM 1/31/2002 +1100, Alan Kent <ajk@mds.rmit.edu.au> wrote:
>On Wed, Jan 30, 2002 at 10:24:26PM +0100, DUBUISSON Olivier wrote:
> > Kerecsen Zsolt wrote:
> > >
> > > Hi,
> > >
> > > I am new in ASN.1, but I know XML.
> > > Do you know anything about the relationship between ASN.1 formal
> > > description and XML Schema definition (XSD)?
> >
> > For what purpose do you need to translate ASN.1 into XSD?
> > An ASN.1 module (when encoded with the XML Encoding Rules of ITU-T Rec.
> > X.693) is an XML Schema.
>
>Could be funky in something like a WSDL file for SOAP.
>Basically there are tools that can handle XSD's but not ASN.1.
>So I can see value in converting an ASN.1 module into a XSD.
>
>Dropping the Z39.50 ASN.1 into SOAP directly would be kind of
>interesting! :-)

There is an example of such an ASN.1 protocol converted into XML
encoding at <http://asf.gils.net/XER>. Here, we used the 1995
version of the Z39.50 protocol. A copy of the specification is at
<http://asf.gils.net/xer/asn1/z3950-1995.txt> and the XML Schema
is at <http://asf.gils.net/xer/ez.xsd>.) Protocol Data Units as
XML instances are at <http://asf.gils.net/xer/pdu_sample.html>.

I believe Matthew Dovey <matthew.dovey@las.ox.ac.uk> used MS
Visual Studio to automatically generate protocol "code stubs"
at the send and receive ends by just identifying the XSD file.
The XSD was generated from the official ASN.1 specification
using "com.wedgetail.asn1.compiler.backends.XMLSchemaEmitter"
by Geoff Elgey <elgey@dstc.qut.edu.au> at the Distributed
Systems Technology Centre in Australia.

This work pre-dates the final version of the ITU-T XER spec,
which I believe is to be available "soon".

The key issue here is really how relevant will be an ASN.1
specification in an XML future. My sense is that technologists
using XML are beginning to see the need for efficient protocols
and some of us believe that ASN.1 has much to offer in that
regard. Even more compelling should be the argument that lots
of hard consensus-building is already represented in the various
existing ASN.1 specifications. A bridge technology like XER
could preserve that hard-won value.

I note with some frustration that some Z39.50 implementors are
working on a "Z39.50 Next Generation" that may employ XML/SOAP
without preserving a strict relationship to the ASN.1 spec.

IMHO, we need to a demonstration HTTP/Z39.50 gateway with a
SOAP listener that translates between Z39.50/XER and Z39.50/BER.
Session parameters above the protocol (host, port, etc.) would
follow the "ZURL" recommendation. In fact, I think it would be
especially useful to have the HTTP/Z39.50 gateway accept a
CGI-style HTTP Post in addition to the SOAP-style HTTP Post.
The CGI-style could implement a more fully articulated ZURL
e.g., <http://www.gils.net/z-url.txt> as is supported in various
Z39.50 tools from Sebastian Hammer and Dave Vieglais, et al.
This also fits in very well with the GILS direction of Microsoft,
<http://ww.gils.net/ms-addins.html> and with the current MS .Net
work on a UDDI/GILS Gateway following Matthew Dovey's prototype.

Received on Thursday, 31 January 2002 06:41:06 UTC