[whatwg] Re: Transition from Legacy to Native rendering

[Ian Hickson]

On Sat, 26 Jun 2004, Jim Ley wrote:
> > >
> > > Oops, it looks like I left out a NOT from the sentance, it's not
> > > flawless, it still turns up both false positives, and false
> > > negatives.
> >
> > Given that to compliant UAs a "conditional comment" is just a comment,
> > any false positive is a bug. How can you get a false negative?
>
> When using it to identify IE, you can use conditional comments to both
> fail to identify IE when it is in use

Could you give an example of this?


> and identify it as IE when it's not

This will only happen if you mis-use them, since it is quite possible to
set up conditional comments so that they appear to just be SGML comments
to other UAs.


> Also please remember you cannot talk about condtional comment
> differences being a compliance issue, conditional comments are not
> outlawed in ECMA 262.

What on Earth has ECMA-262 got to do with anything here?


> > > I also don't really see why the fact it violates the spec here is a bad
> > > thing, but adding innerHTML violations are okay -
> >
> > Who said they were ok?
>
> Oh right,  so you think I'll have an easy job getting it removed from
> Mozilla then as it's non-conformant?

I doubt it. Real world Web sites depend on those extensions.


> If not, and it's fine for "conformant UA's" to include that, then it's
> just as fine for everyone else to include other logical extensions.

I am baffled why, if you think it's fine for everyone to include logical
extensions, why you are "disgusted with Opera, Apples and Mozilla
involvement in this deliberate subversion of internet standards".


> > No, they're not. They all get returned as strings, not at all
> > following the principles outlined in the section you cited.
>
> Since they're not validating UA's there's no doctype reading going on,
> in any case, my example page didn't provide a DTD, and so string was a
> perfectly acceptable choice IMO.

Not a forwards-compatible choice if that attribute was defined as a
boolean or a number, though, is it.

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