- From: Ian Hickson <ian@hixie.ch>
- Date: Fri, 9 Jun 2006 01:49:50 +0000 (UTC)
On Thu, 8 Jun 2006, L. David Baron wrote:
>
> On Friday 2006-06-09 01:17 +0000, Ian Hickson wrote:
> > Given that requiring a new flag per feature is not an option (as, as
> > mentioned before, it would require a central authority to add these
> > features, slowing the introduction of new features and discouraging
> > experimentation), what solution would you propose instead?
>
> I think it is an option, and I don't see why you're so insistent that it
> isn't. Authors are actually going to want interoperability; to get
> that, it's required.
I don't think it's an option because:
<input type="text" required maxlength="80" name="subject"
spellcheck list="subjects" value="Hey there"
inputmode="user startUpper">
...or:
<input type="text" required name="formula" spellcheck="math"
inputmode="math" highlight="math" auto-evaluate>
...or:
<textarea name="source" rows="80" spellcheck="off" autoindent="C++"
highlight="C++" auto-evaluate="off">
...gets out of hand very fast.
<input> already has 27 element-specific attributes, plus 5 global
attributes, plus 1 deprecated attribute, plus an uncountable number of
event handler attributes. Going the road of one-attribute-per-feature
would escalate that even faster, especially given that these features
don't affect anything other than the way the user interacts with the
input field (i.e. they don't affect the actual allowed values, etc).
(Also, apparently Mozilla's implementors for this feature have received a
request that the feature validate per HTML4's DTD. While I don't consider
this a sensible request, and it doesn't affect my opinion of how the
feature should be added, it is worth noting that accept="" on <input>
elements does current validate per the HTML4 DTD.)
--
Ian Hickson U+1047E )\._.,--....,'``. fL
http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,.
Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Thursday, 8 June 2006 18:49:50 UTC