New homepage design of d-p-l.org is now live. <eom>

Fri Dec 16 19:10:37 PST 2011

On Friday, 16 December 2011 at 23:41:19 UTC, Stewart Gordon wrote:
> Yes:

I'll agree that some of the validator's things help
with that, but not all of it.

If you write <a href="#"><div>block in inline</div></a>,
the validator will reject it, but it works... and that's
a useful thing when doing drag+drop applications (since
older IE doesn't let you drag other elements).

Though, I can see your point with cross-browser
compatiblity, to an extent, as that code sometimes (not
always... it can change across refreshes of the same
page...) brings out bugs in Firefox 3.6.

Anyway, though, the specific doctype still isn't terribly
important, since, in practice, tools tend to ignore it
anyway. Browsers see it's presence as an on/off switch
with standards compliance mode vs quirks mode; declaring
the wrong one doesn't break anything. (Indeed, HTML5
has agreed to use the common, previously wrong, shorthand
of <!DOCTYPE html> as the new standard!)

Stuff like improperly closed tags or bad entity
encoding can break, but that's pretty well independent
of doctype validation. That's simply a matter of the
document being well-formed.

> What are these "things that are useful" to which you refer?

There's the drag and drop issue from above, the custom
attributes thing from below, and sometimes, using certain tags
or generally accepted shorthand. (For instance, <script>
used to require a type, but it worked without it anyway.
Again, the html5 folks decided to adjust the standard to
fit the practice - something I actually like about them. This
is a really minor thing, though.)

> Are custom attributes distinguished from standard attributes in 
> some way, in order to keep attributes invented by different 
> browser manufacturers from clashing with each other and with 
> attributes that become part of a later HTML standard?

They are in html5 - the data- prefix is allowed and reserved
for the user. In older versions of the html standard, there
weren't allowed at all, whether prefixed or not. (They did
work in practice, though.)

I like custom attributes a lot, since they add a richness
that Javascript (and CSS too) can exploit in interactive
pages.

They're the main thing I miss if I validate with one of the
other DTDs.