W3C home > Mailing lists > Public > www-html@w3.org > May 2007

Re: Interoperability/Compatibility Re: Semicolon after entities

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Tue, 1 May 2007 21:03:06 +0300 (EEST)
To: www-html@w3.org
Message-ID: <Pine.GSO.4.64.0705012048210.27896@hopeatilhi.cs.tut.fi>

On Tue, 1 May 2007, Karl Dubost wrote:

> Le 30 avr. 2007 à 22:02, Lachlan Hunt a écrit :
>>> Second, what you really mean is working the same way as other browsers.
>> 
>> Yes, that's what interoperability means.
>
> Just to fix a common error in the vocabulary.
> This is *not* what interoperability means.

It seems obvious to me that interoperability does not mean identical 
behavior, but it is less obvious what it really means.

> Interoperability means the capability of two independent implementations to 
> work exactly the same.

I don't think so.

> 	A document is read or created by
>       applications foo and bar developed
>       without prior knowledge of the other
> 	application.

I don't know what to think about that. It sounds like a requirement on 
independent implementations, nothing more, nothing less.

> What is described: "working the same way as other browsers" is compatibility. 
> It is completely different.

I'd say it's similarity.

Merriam-Webster defines interoperability as follows:
"ability of a system (as a weapons system) to work with or use the parts 
or equipment of another system"
   http://www.m-w.com/dictionary/interoperability

There's some military emphasis, so let's have a look at a specifically 
Internet-related definition, in RFC 4775:

"that multiple products implementing a standard are able to work together 
in order to deliver valuable functions to the Internet's users".

This postulates
a) the existence of a "standard" of some kind - a specification that is
    implemented in some products
b) products "working together" (which is where "interoperability" comes
    from, etymologically - well, it means "working between (some
    entities)") in a useful way.

Thus, interoperability of browsers is irrelevant, not an applicable 
concept, since browsers are not required to work together and they don't 
work together. Instead, a browser must be interoperable with _servers_ (as 
well as possibly other software, such as plugins - but hardly other 
browsers).

When two browsers render a document according to applicable HTML and other 
specifications and in a useful way, they could be said "compatible" with 
each other, but that's really an unnecessary concept. What matters is that 
a browser complies with specifications; this would matter even if there 
were just a single browser in the world. If the specifications are 
reasonable, two browsers that comply with them are necessarily 
"compatible" in the sense of performing the same job properly. But it 
would be absurd to require that they behave identically, still less that 
they are "compatible" or "interoperable" with some dated browsers that 
violate the specifications.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Tuesday, 1 May 2007 18:03:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:16:10 GMT