Re: relevance of diverse HTML authoring practices [was: Versioning re-visited ...]

On Jun 23, 2007, at 10:04 AM, Philip & Le Khanh wrote:

> Dan Connolly wrote:
> > Speaking personally, my goal is that the specification we
> > deliver is sufficient that new implementations based on it
> > will interoperate with a critical mass of the deployed content
> > on the web. Much (most?) of that content reflects
> > poor authoring practices. These practices are very much
> > relevant considerations for our design decisions.
> Dan, I think you are trying to hit two nails with one hammer.
> By all means let there be a specification for /browsers/ that
> can handle the majority of extant content : but let that
> specification not influence the design of HTML 5, which I
> (for one) see as the primary task with which this group is
> concerned.  To put it another way, let the specification
> for HTML 5 inform the specification for browsers; but
> /please/ do in any way allow the specification for browsers
> to inform the specification for HTML 5.

If I understand your proposal correctly, it is to have a  
specification for what browsers should implement, and then a separate  
real HTML5 that is completely uninformed by the browser  
specification. It seems to me, then, that this second specification  
would be an excercise in pointlessness, since by design browsers  
would not support it, and therefore content would have no reason to  
follow it.

> Or to put it more simply, we must not allow the tail to wag the dog.

The way I think of it, the living, existing World Wide Web is the dog  
and written specs about how it should work are the tail. I imagine  
you may be thinking of it the other way around.


Received on Sunday, 24 June 2007 23:47:30 UTC