[whatwg] [html5] tags, elements and generated DOM

[Jim Ley]

On Apr 7, 2005 10:24 AM, Olav Junker Kjær <olav at olav.dk> wrote:
> Jim Ley wrote:
> > Firstly I think the conclusions that the audience for WHAT-WG stuff
> > doesn't understand the limitations of the validator is sustainable -
> > where's the evidence?
> 
> People putting small icons on their pages to indicate that the page is
> valid. Also, lots of articles on the web about jumping through hoops to
> e.g. make a flash embed validate.

Which doesn't say anything that these users believe anything more of
HTML validation than it is, it's a very important _part_ of QA. Given
that there are no complete HTML conformance checkers in existence
today for existing HTML technologies.  It seems very strange to remove
one of the few parts of QA available so what have we then got.  Or are
the WHAT-WG members going to step up and implement one?

> As HTML applications becomes more complex

I thought the whole point of the WHAT work was to make HTML
applications simpler, not more complex, are you suggesting the current
specs are failing in this area?

> However, a
> syntax error in the initial value of a date control *will* cause the
> page to stop working as intended.

Could you describe how?  My reading of the error handling defined in
the spec for that situation does not lead to the failure you describe.
 However the unclosed <B> element does exactly that. (in the XHTML
dialect)

> The reason comparatively few authors validate their pages, is that the
> immediate practical benefit is quite small. 

Certainly, but the point is that it is valuable as part of a larger
automated QA process, most people rely on non-automated QA methods,
that's fine, if you can find and afford the testers.

> A conformance checker would
> be much more valuable since it might catch real errors which might cause
> the page to stop working. 

But who's going to write it?  There's no point talking about perfect
tools when no-one's writing it...

Jim.