RE: Call For Feedback: SIDL Proposal

Actually, XMI isn't that great for IDL.  When I was at IBM, I architected a
solution using XMI 1.1 to describe SOAP messages.  I called it XIDL and
presented it - among other things - at XTech 2000.  XMI suffers from many
shortcomings, such as lack of namespaces, poor reference/link support,
arcane attribute naming style ("fieldN...."), lack of tooling, verboseness,
and extreeme difficulty of humans to author XMI documents.  People who think
Schema is hard to author and verbose will have a difficult time with XMI.

Cheers,
Dave Orchard
Lead Architect
Jamcracker, Inc.
935 Stewart Dr.
Sunnyvale, CA 94086
p: 408.830.1886
f: 408.328.0936

Named to Red Herring's list of 100 Most Important Companies:
www.redherring.com/mag/issue79/herring100/jamcracker.html

Named to Fortune's list of Cool Companies 2000:
http://www.fortune.com/fortune/cool/coo.html

-----Original Message-----
From: xml-dist-app-request@w3.org [mailto:xml-dist-app-request@w3.org]On
Behalf Of cap@novosoft.nsc.ru
Sent: Tuesday, July 11, 2000 5:41 AM
To: Henry Lowe
Cc: xml-dist-app@w3.org
Subject: Re: Call For Feedback: SIDL Proposal



07/11/2000 07:06:24 xml-dist-app-request wrote:
>Hi,
>
>May I suggest that instead of inventing a new interface
>description grammar, that an open, well established
>IDL is used for SOAP, XML-RPC, etc.  ISO/IEC 14750 has
>been a standard for about 2 years now, is implemented
>in large number of commercially supported products,
>is used as a language independent interface for testing,
>is the CORBA IDL and is easily mapped into the DCE
>RPC IDL (a demonstration it can support a full RPC in
>addition to OO).
>
>Sorry, but I must have missed James message last month.
>
My mail clent crashed on the first try to send the reply,
so this may be duplicate mail with a bit different content.

1. I do not like ISO process orgainsation. They sell standards.
And it is diffcult to look into standard and check whether
an implementation is conforming or not.

OMG, ECMA, W3C, JavaSoft policy on it is much better.
I hardly can call ISO process open for consumers of
technology. It also will bring extra cost that could be
unacceptable to following groups:
- open source  developers
- students (that are usually good source of feedback
  and early adopters of interesting technologies (they
  are also future professionals, but it is another story)).
- companies in developing countries

2. I do not know what this particular standard is about.
Is it similar CORBA IDL or what?

If yes, I will not support your suggestion. Getting parser
implemented right always were major problem in the open
source CORBA implementations. The preprocessor+parser
scheme was difficult to implement with conventional parser
generators and is the big source of bugs.

3. I would suggest to use OMG XMI 1.1 as foundation for IDL
language. The second pat of the standard, has good examples
how it could be done.

The IDL may be expressed as UML profile or completly
separate application of XMI 1.1. Provision of formal
metamodel for XML-RPC IDL, will also help understanding
of of the specification. I think that MOF 1.3 specification
may be a good example of how this could be done.

XMI 1.1
ftp://ftp.omg.org/pub/docs/ad/99-10-03.pdf
ftp://ftp.omg.org/pub/docs/ad/99-10-02.pdf

MOF 1.3
ftp://ftp.omg.org/pub/docs/formal/00-04-03.pdf

Constantine

Received on Tuesday, 11 July 2000 13:39:40 UTC