Postel and ESR (was Re: [TriLUG] Monthly mailman reminders and procmail)

Mike M no-linux-support at earthlink.net
Sat Oct 2 11:10:41 EDT 2004


On Fri, Oct 01, 2004 at 08:59:53PM -0400, Aaron S. Joyner wrote:
> Mike M wrote:
> 
> >Nevertheless HTML is robust.
> >
> Tried to make moderately complicated cross-browser compatible javascript 
> before?  Even some pure HTML constructs simply don't fly with all of the 
> browsers.  Hence the problem ESR was alluding too.  

That would be the JavaScript engine's and the browser's fault from not 
applying Postel's Law.

Postel's Law works amoung entities that are trying to cooperate.
ESR's questioning of Postel's Law based on an episode in the long
running Browser War is flawed because in this arena Postel's Law is not 
being used intentionally. ESR himself does not apply Postel's Law which
is plainly evident in his large and strongly opinionated body of work.
In a discussion about browsers not cooperating ESR attacks a respected
and proven rule on how to cooperate. ESR is just not a credible authority 
on Postel's Law and is unfit to modify it. Postel's Law stands.

Since the browser makers refuse to apply Postel's Law, users such as
ourselves apply it for them.  We liberally use Gecko technology, IE,
ActiveX, JavaScript, etc. to consume the web.  When we provide web content
we use tags that are known to work with the greatest number of browsers.

> Just to clarify, I 
> do think that generally HTML works remarkably well considering all that 
> is asked of it these days. 

Aye. Thanks to lot's of people using Postel's Law.

> I just always associate ESR's account of 
> McIlroy's take on Postel, and the occasional little personal horrors 
> I've encountered in dealing with HTML.  :)

Perhaps your situation is like being a humanitarian aid worker in Iraq.
You are trying to do some good and party's on all sides are making
your job more difficult.  I wonder what Rodney King would say?

-- 
Mike

Moving forward in pushing back the envelope of the corporate paradigm.



More information about the TriLUG mailing list