Re: Unexpected valid CSS

On 12-Sep-08, at 6:30 AM, Paul Arzul wrote:
> should we be ignoring obvious typographical errors
> ("ody" was surely meant to be "body")? isn't this what warnings are
> for? further, if the css was submitted/found via (x)html, then the
> validator *knows* that this is an error as it is aware of the context.

I don't think there is any reasonable way to "ignore" typos. A  
checking tool should not second-guess what the author did and omit to  
output some warnings or errors because "maybe that was on purpose".  
However, the only thing that stops a checker like the validator to  
improve from
"Property colour doesn't exist"
to
"Property colour doesn't exist. Did you mean “color”?"
is a good patch.

> a lookup table of valid css properties or (x)html elements
> cross-referenced via the edit distance[1,2] algorithm would be rather
> helpful, imho.

Agreed. Now, honestly, as much as I'd like to I can't go to any of our  
staff and suggest that they could implement this. I won't go into  
details, but we're already struggling to push a number of crucial  
patches on CSS grammar, CSS3 updates, etc. out of the door. The only  
way a feature such as the distance-check could be developed soon is  
through community patches.

Let me extend, again, an invitation to java hackers:
* the code is here: http://dev.w3.org/cvsweb/2002/css-validator/
* we have a staging server to test development and patches
* we'll happily give commit rights to developers ready to help, hack,  
improve

In the meantime, I'll encourage you to enter the suggestion in bugzilla:
http://www.w3.org/Bugs/Public/enter_bug.cgi?product=CSSValidator

Thanks,
olivier
-- 
olivier Thereaux - W3C - http://www.w3.org/People/olivier
W3C Open Source Software : http://www.w3.org/Status