W3C home > Mailing lists > Public > www-tag@w3.org > December 2008

Interoperability vs. Uniformity

From: Larry Masinter <masinter@adobe.com>
Date: Tue, 30 Dec 2008 05:29:54 -0800
To: "www-tag@w3.org" <www-tag@w3.org>
Message-ID: <8B62A039C620904E92F1233570534C9B0118C8161C59@nambx04.corp.adobe.com>
In several of the discussions, the word "interoperability" has been used to describe something that I believe would be more accurately called "uniformity". Two components "interoperate" if they "operate" when connected.  A browser and a web server "interoperate" if, when the browser sends something to the web server, which sends something back, the web server's response is sensible (it doesn't crash or send back garbage) and the browser does something sensible with that response.

There is no requirement that two different web servers respond in the same way to similar requests, nor is there a requirement that two different browsers behave similarly, in order to say there is "interoperability".

Interoperability is generally desirable, in the same way that "sensible" behavior is.

On the other hand, two different browsers or web servers are "uniform" if their behavior is similar  -- if both servers implement the same options for "https" and accept the same login parameters, or if both browsers display the same ill-matched HTML document in the same way.

Uniformity may or may not be desirable, depending on your perspective. Certainly uniformity is desirable from the point of view of understanding what will likely happen. Uniformity is also often at odds with extensibility and innovation: if everything is required to be the same, how can anyone invent something new?

Larry
--
http://larry.masinter.net
Received on Tuesday, 30 December 2008 13:30:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:48:09 GMT