Re: An HTML language specification vs. a browser specification

On Wed, 19 Nov 2008, Julian Reschke wrote:
> > 
> > I certainly would hope that if I started writing complete gibberish in 
> > the HTML5 spec, that the working group would oust me. Indeed, I am 
> > relying on the fact that the working group has _not_ expelled me as 
> > evidence of continued overall support for my work. If I'm doing a bad 
> > job, or if any editor is doing a bad job, then the working group had 
> > better retract their support, or the Web as a whole will suffer.
> > 
> > Do you disagree?
> 
> No.

Ok, great. Then I don't think there is any practical disagreement here.


> But that's not the same thing as "as good as Ian"

While I certainly wouldn't have phrased it that way myself, I'm curious as 
to whether you are concerned that that would be leaving the bar too low or 
too high.


> or "the same way as Ian is doing it".

I agree that we shouldn't expect people to use the same style as I do. 
There are many ways to write a high quality modern specification, and 
indeed my own style is a hodgepodge of different techniques, including for 
example a lot of algorithmic/imperative style requirements with explicit 
steps, some state-machine like requirements, and some constraint-like 
requirements. That said, I _would_ expect that the working group would 
hold any editor to the same level of quality in terms of making sure that 
all error handling is defined, that ambiguities don't exist where they can 
affect interoperability, that research and data is held to a high regard, 
that use cases trump theoretical purity, and so forth. Basically, 
following our design principles.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Wednesday, 19 November 2008 08:49:33 UTC