Re: Cleaning House

mjs@apple.com (Maciej Stachowiak) wrote:

>I think bridging the descriptivist/prescriptivist philosophical chasm is
>probably beyond the scope of this mailing list and thus not worth arguing
>about much further. It took decades to settle in the field of linguistics
>and we're not going to settle it quickly in the field of web standards.
>It's good to recognize the difference, but I can't see either side
>convincing the other.

The two are not inherently mutually exclusive.

Your perspective tends to the Descriptive because you spend all 
day trying to deal sensibly with the language in the wild. Your 
experience informs us that there is a need to take into account 
the actual permutations found in usage.

However, uncontrolled mutations have unpredictable consequences 
so there is a need for direction which can only be provided by a 
Prescriptive approach.

At least in IETF standards processes there is a common 
understanding that one needs to both document existing practice 
_and_ define the correct way forward.


In the context of evolving HTML, that means there is a 
legitimate need to define how to deal with existing content — 
which, BTW, includes not only “tagsoup”, but also content 
written to be in de jure compliance with previous specifications 
in the field — but also to specify the correct way to do 
things today with an intent to move things forward to how they 
should be done in the future.

We cannot possibly hope to achieve it all in one step so 
compromises need be made, but we cannot abandon any progress 
toward improvement merely because previous attempts have been 
met with arguably limited success.


Without arguing the merits of any one specific issue (IOW, don't 
assume I agree on a particular point just because I use it as an 
example), the current WHAT WG submission describes both a 
“tagsoup parser” (no negative connotation intended) syntax 
and a XML-based format. The former specifies how to deal with 
HTML “in the wild” and documents existing parsers in 
deployed desktop browsers; the latter provides a way toward the 
goals for the future.

One could easily (and I use the word with a certain amount of 
irony) produce a specification that satisfied both goals.

One point of discussion would be whether the language was 
defined by its strict XML-based syntax and pared-down 
vocabulary, and with the specification for handling legacy 
content relegated to “mere” documentation of existing 
practice; or whether the specification was the legacy one with 
the strict version a mere idealized subset.

I think HTML 4.01's Strict and Transitional distinction might 
inform us in that debate, though I suspect arguments can — and 
probably will — be made on all sides citing it.


However, I don't think we need to “bridge the 
descriptivist/prescriptivist philosophical chasm” in any 
context other than communication, so we may better understand 
each others' arguments. A specification such as we are tasked to 
produce is by its very nature Prescriptive; the question is how 
much Descriptivism should be applied in its making.



-- 
“I don't want to learn to manage my anger; I want to FRANCHISE it!”
                                                     -- Kevin Martin

Received on Sunday, 6 May 2007 12:25:46 UTC