Some will. Some won't. Some people will spend a lot of time undoing the
mistakes they learned from someone who considered it "mildly irritating" to
scroll to the top and declare a new style or class.
>Maybe, but I would expect direct styling to be used so infrequently that
>the novice would be unlikely to be confused.
I would be surprised if this is the case. If you offer people a shortcut.
They will take it. That's human nature.
>Extra features *can* make it harder to use a language. I think they
>*must* make it harder to use a language like a native, but not every
>language is structured in a way that makes learning a subset
>impractical.
I agree. HTML is nicely segmented now, so that style sheets and HTML itself
are clearly divided. Direct style application would blur this distinction
in user's minds (as I described), and in the implementation (as Glenn
described).
>So, how do you feel about naked lambdas in LISP (a language which does a
>good job of not forcing you to name things unless they need names)?
>Do you think LISP would be a "better" language if every function had to
>be created with defun?
I feel that lambdas (naked or named) are an integral part of LISP and should
be well integrated into the LISP language. Styles, (named or not) are _not_
an integral part of HTML and they should stay in their own style sheets or
in their STYLE element. You seem to be forgetting this. HTML is not a
style sheet language and style sheet directives would ideally not appear in
HTML documents at all. The fact that they can appear in the STYLE element
is a compromise to make editing easier. Just how much farther do we have to
go? Back to <CENTER> and <BLINK>?
I grudgingly support the STYLE attribute using named styles, because it is
better that people abuse the STYLE attribute (which they will) than the
CLASS attribute. If we get rid of STYLE, people will say "CLASS=BLUE" which
is worse than if the element didn't have a CLASS at all.
Paul Prescod