W3C home > Mailing lists > Public > www-tag@w3.org > February 2003

RE: Precise Definition for Interoperability Needed (Was RE:[Minut es] 6-7 Feb 2003 TAG ftf meeting (why XML))

From: Bullard, Claude L (Len) <clbullar@ingr.com>
Date: Mon, 17 Feb 2003 09:02:54 -0600
Message-ID: <15725CF6AFE2F34DB8A5B4770B7334EEEACF28@hq1.pcmail.ingr.com>
To: "'Rick Jelliffe'" <ricko@topologi.com>, Chris Lilley <chris@w3.org>
Cc: www-tag@w3.org

I've tried to distinguish portability from
interoperability.   Portability means the 
systems agree to XML per XML 1.0, so to the 
point of parsing and validating, they can 
operate.  Interoperability means they agree 
on the semantics of the message or document, 
that is, as in ontological commitment, will 
not in response, exhibit an observable behavior that 
contradicts the agreements.  While narrow 
and sometimes not achievable 100%, this 
definition for interoperability better 
describes what people think XML can do 
for them but requires more than a simple 
well-formed definition.  Note also by 
that definition, there can be percentages 
of interoperability although I have no 
particular measure in mind.


From: Rick Jelliffe [mailto:ricko@topologi.com]

Dan Connolly wrote:

> Please direct your suggestions to Chris
> in particular (with copy to www-tag) as
> he has the action to do the next draft for review
> of section 3 on formats.

When discussing "interoperability", it might be useful to contrast "guaranteed interoperability"
--i.e. where the receiver can always accept the data correctly-- with  "robustness"
--i.e. where the receiver will always fail if it cannot accept the data correctly--.

These are in distinct from "unreliablility".  XML has never had guaranteed interoperability
but it has had robustness.   E.g. an XML processor is not required to parse ISO 8859-1
documents, but is supposed to fail.

Some people just need "robustness": for example, people making data available on the
WWW in the most convenient form for the sender. 

Others may need "guaranteed interoperability", but niche-users can get this now by 
profiling XML.  So I suspect the emphasis should be robustness as the bottom line.
Received on Monday, 17 February 2003 10:03:33 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:37 UTC