W3C home > Mailing lists > Public > www-tag@w3.org > January 2009

Re: Extensibility and Uniformity; defining "error" behavior

From: Rigo Wenning <rigo@w3.org>
Date: Fri, 9 Jan 2009 10:40:43 +0100
To: "Ray Denenberg, Library of Congress" <rden@loc.gov>
Cc: www-tag@w3.org
Message-Id: <200901091040.44287.rigo@w3.org>

"renaming game" was not intended to mean playing games here. Sorry for 
being imprecise in my language, there was no offense intended.

On Tuesday 06 January 2009, Ray Denenberg, Library of Congress wrote:
> It isn't my intent to prolonge this debate, I simply would like to
> register my dissent over defining interoperability as "the ability
> to pass certain tests with a well defined result".
>  "Interoperability"  is an important and well-understood word in
> standards, and mis-using it will cause confusion.

Since more than 2 years, I'm part of a group tasked by the European 
Commission to accompany the re-evaluation of the European 
Standardization landscape. The group got through the exercise of 
trying to understand the various notions, meanings and layers 
of "interoperability".

If even those guys that are 30 years in "standardization" try to 
understand the meaning of "interoperability", I do not think 
that "Interoperability is [...] a well understood word."

The group identified several layers:
There is technical interoperability (also interoperation, yes, the 
power-plug nightmare in Europe). There is syntactical 
interoperability (I think we are talking about it here) where two 
engines are supposed to produce the same result from the same input. 
There is semantic interoperability, where people have the same 
assumption/definition on the terms they are exchanging. And finally, 
there is procedural interoperability (ISO 9000, eGov etc). And I do 
not get all the edge cases with those definitions.

Now we have "interoperability" vs "uniformity" where the discussion is 
centered around the level of detail that goes into a specification. 
There we have also a scaling from a pure interoperation with few 
details in the specification up to the "reference code model" where 
the code replaces the specification. This is a new aspect that I 
learned here and that I will bring back to the group in Brussels. 

Note that from my legal background I'm reluctant to use detailed terms 
as they preclude evolution by new interpretations of known concepts. 
At least I would really like detailed definitions to be sandboxed, as 
is the case for CSS 2.1. IMHO.



Received on Friday, 9 January 2009 09:41:19 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:26 UTC