Re: Cleaning House

On May 6, 2007, at 4:47 AM, Tina Holmboe wrote:

> On Sun, May 06, 2007 at 04:21:24AM -0700, Maciej Stachowiak wrote:
>
>> So there you go. I think bridging the descriptivist/prescriptivist
>> philosophical chasm is probably beyond the scope of this mailing list
>
>   If, by "descriptivist philosophy" applied to web standards, you
>   mean "create a new standard by documenting what people do" - then  
> yes,
>   I think it impossible to bridge.

My own personal stance would be less extreme than this, merely that  
new versions of the standard should be informed by how previous  
versions were actually used, and adapt some extent. Markup languages  
are not quite as fluid as human languages. But they are more fluid  
than, say, strongly typed imperative programming languages, and they  
have to be, since they mix with human language.

>   But I'd love to see anyone write a serious dictionary by simply
>   and mechanically taking down whatever people say and then suggest
>   everyone else should use it to learn to communicate ...

Your request is granted: see the Oxford English Dictionary.  
Everything in the OED is based on print citations showing words being  
used in a particular sense: <http://www.oed.com/about/>. It is one of  
the most widely respected dictionaries of the English Language.

>   Just please don't expect anyone to actually grasp that my
>   example
>
>     <b style="position: absolute ; top: 150px;">Welcome!</b>
>
>   is an actual header just because the author thought it was
>   a good idea at the time.

This probably would be non-trivial to deduce, yes, but I also think  
this is a pretty rare way to say "header" compared to <div  
class="header"> or <h1>.

>   PS: half an hour of explanation and he was using EM and
>   H*. It's a matter of knowledge. Imagine what the new WA1
>   draft will do to these authors?
>
>   If accepted, I can only shudder to think how we, that is those
>   of us "in the trenches", are supposed to explain the
>   situation.

I don't see how it does anything to encourage use of markup like the  
above to represent a header. It includes <header> and <h*> elements  
for headers containing blocks or only inline-level content; it  
currently makes the 'style' attribute non-conforming on most  
elements; and it still has <em> and <strong>. It also defines an  
algorithm for mapping sections to corresponding headers, giving a  
potential practical benefit to using more semantically rich elements  
for headers, if UAs choose to expose section/header navigation.

It seems like part of your objection may be based on unfamiliarity  
with the contents of the spec.

Regards,
Maciej

Received on Sunday, 6 May 2007 12:06:28 UTC