Re: CSS3 module: W3C selectors

[[Disclaimer: I am not commenting negatively on the work of the selector
proposal working group; I am saddened by the general direction I perceive
that CSS is taking, and that I can only point at by pointing at specific
aspects of specific proposals.]]


Chris Lilley <chris@w3.org> wrote:
>To put that question in its correct chronological order,
>why was XPath such a departure from W3C selectors? 

The relevant question is: should one adopt a mediocre solution?

The question is not about what was first, but what is better.  If XPath is
good for CSS purposes, by all means let's adopt it, if it is not, don't
let's.  As I argued in
http://lists.w3.org/Archives/Public/www-style/2001Feb/0000.html, it would be
a bad thing to adopt the current CCS3 selector proposal.

The second example I gave there shows a language that does the trick, and in
a way that non-CS people can understand easily.  (That example is not a
proposal, but a proof that it can be done.)


CSS started out as a great solution, but what's happening with it is making
me sad.  It seems to be deteriorating into another bit of bloatware that
will require knowledge of hundreds of reserved words to understand.
A good language allows one to think clearly and thereby know what to write,
instead of forcing one to carry a heavy reference volume around.
A good language allows to express what the user has in mind, not just what
the designer had in mind.
A good language makes the obvious easy, and the thinkable possible.
A good language implements a small set of principles that fit naturally to
the domain and that can be combined in an intuitive way to express what one
wants to express.

If it were just the selector language, my comments might be seen as an
overreaction.  But it is happening at many places in the proposals, e.g. see
my http://lists.w3.org/Archives/Public/www-style/2001Jan/0073.html for
another example.

J. A. Durieux
http://www.biep.org/

Received on Friday, 16 February 2001 06:58:57 UTC