Re: DogFood (and inline/block constraints)

I'm not copying public-html until I find time
to cite earlier discussion of this issue...

On Thu, 2007-12-06 at 00:15 +0200, Henri Sivonen wrote:
> On Dec 5, 2007, at 21:29, Sam Ruby wrote:
> 
> > But this is clearly just the start.  At the present time,  
> > html5.validator.nu conformance checking tool identifies a gross  
> > number of errors:
> >
> >  http://tinyurl.com/2wywpp
> >
> > I imagine that the errors can be grouped into three categories:
> >
> >  1) errors for which my page ought to be fixed
> >  2) errors for which the validator ought to be fixed
> 
> Did you identify any errors that the validator reported but were not  
> errors per current spec draft? Granted, there are a number of  
> unfriendly messages there, but as far as I can tell, all of those  
> point out actual errors.
> 
> Even "Bad value text for attribute type on XHTML element input." is  
> technically correct although not exactly helpful (see error class 4  
> below). This message is the #1 usability bug, BTW.

I think it's a spec bug.

I don't think it's cost-effective to try to constrain
authors in this respect. I hope this constraint is
dropped from HTML 5.


> >  3) errors for which the spec ought to be fixed
> >
> > I'm posting my early results here in the off chance that it sparks a  
> > discussion.
> 
> 
> I identified four classes of errors:
>   1) meta charset in XHTML
>   2) wbr
>   3) a dangling for="q" attribute
>   4) lots and *lots* of cases where you have inline content where only  
> block content is conforming.
> 
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
gpg D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E