Re: Class Selectors Extension Proposals

On Fri, 17 May 2002 09:45:56 +0200, you wrote:

>Jan Roland Eriksson wrote:
>
> > It was supposed to be "typography on the web" from the start of it,
> > (words said to me from at least one of the original creators of CSS)
> > why is it necessary to molest that original idea?
>
>For two good reasons :
>
>1) only stupid people never ever change of opinion

I don't really know if I shall say " true " or " thanks :-( "

>2) why not ? Better stuff always comes from molesting the old ideas.

Sure, but then you are also creating a "new" product that deserves a
"new" name.

>   Molesting the original idea of CSS, I was able to make a
>   transformation language, that at least two of the original creators
>   of CSS found very interesting ;-) That was not in the "line" of
>   CSS and the temple's keepers never thought about it, SO WHAT ?

You still seem to have problem to understand my distinction.

CSS is defined as a "descriptive only" language, much in the same way
as traditional markup is descriptive only.

Per definition a descriptive language can never "run" any other code
on a users system than what the users client is already programmed to
do with the description it receives, i.e. a language that only
describes a suggested behavior can never introduce a security "black
hole" client side.

Now, start to add conditional evaluation of expressions in a language
and you are directly opening up for "code kiddies" to send their
unproven code down to users to be executed on the users system.

We have had enough of bad scripting on the internet already, I have
people at my current work site that sometimes calls for "support" from
every "black daemon" they can think of at times when they can not get
out of some site they have "been shoved" into, usually by mistaken
"click" at some point.

A "long" time ago even Pei Wei [1] realized that he had given birth
to an "angel with thangs" and that was in the early 90'ies already, in
Viola. Of course that did not stop Netscape from introducing
"Javascript" later on, for all the agony that has caused.

> >> Anyway, I think it deserves to be added to the CSS Suggestions
> >> List...

No; to a list of some new yet unknown presentational language maybe,
but not to CSS.

> > Do _not_ add any more types of "procedure/function calls" to CSS than
> > what is already there in CSS2. Instead spend serious time on that
> > "promised" CSS2.1 spec. The world of users and authors needs it,
> > implementors needs it, don't you?
>
>I am an implementor, an author and an end-user. And I simply have
>opinions on stylesheets that you don't share. That's not a secret.

Of course not, I just wish for you to see that borderlines needs to be
drawn. Merging all kinds of presentational ideas into a "one size fits
all" model is not going to be a good thing.

CSS is a descriptive language only; massage your own, use the syntax
if you see that as a benefit, but give it a new name and go from there
to convince implementors to do things your way.

[...]

> > Or is "constant earthquake" the real future?
>
>The only constant earthquake I see here is your aggressiveness when
>you answer to a technical proposal that is not in your "right line".

My "aggressiveness" comes from the inability of several participants
here to see that there is (and shall be) a definite borderline between
a style language that can only make use of fixed pre-programmed
routines in a user agent, and a style language that calls for a real
"conditional evaluation processor" on the users side.

CSS shall be self sufficient in the "fixed pre-programmed"
environment, other variations of "transformation languages" shall be
named differently.

[1] Pei Y. Wei, the creator of the Viola browser and the
    original inventor of client side executable code as
    in a scripting language. http://www.viola.org/


-- 
Rex [the fox in the chicken shack]

Received on Friday, 24 May 2002 20:18:22 UTC