Re: "maybe even in the fact that you use words as all," (sic)

On Thursday 17 July 2008 19:10, Tab Atkins Jr. wrote:

> While a wonderful dream, you're effectively asking for machine
> translation into a vastly varying medium, while machine translation
> between ordinary languages is still an extremely imperfect process. 
> No amount of xml or tagging will do this appropriately even if you
> *could* convince authors to integrate it into their markup (which
> they won't, in general, as even a flawed effort would require a
> massive expenditure of effort to achieve anything close to
> reasonable); you need strong AI that can read and understand a
> document by itself and then intelligently translate it into another
> medium.
>
> So, in order to do this right, we need strong AI, and once we have
> strong AI, we don't need to expend extra effort on doing this.  So
> the correct course of action is to do nothing in this regard and
> instead fund AI researchers.  ^_^
>
> We can still try to improve our accessibility, of course, in ways
> that *don't* require enormous expenditures of effort to achieve
> something useful (that's one part of semantic markup, after all), but
> not in the directions that you're discussing.

The Web architecture that W3C is developing with its various 
specifications is meant to handle the long, medium and short term all 
at once. (The result therefore is not always as elegant as one may 
wish...) In the long term, computers will no doubt be able to do more 
on their own: find hidden information and combine and reformat it for 
us. In the short term, computers need help from the authors. But if 
that is extra effort for the author, the effort has to be made 
attractive somehow.

As an example, POSH (plain old semantic HTML) is easier to adapt to 
different media than font tags and tables. But many authors don't care 
about other media. However, they do appreciate that HTML + CSS allows 
more style than HTML on its own and makes their downloads quicker, too.

You can see this process at work in many W3C standards. They have to be 
stepping stones on the way to the semantic Web (which, like its 
complement, AI, is a direction, more than a location), but if we want 
people to actually step in that direction, the standards must bring 
short-term gains as well.

Of course, "the" author doesn't exist. There are many different people. 
One size doesn't fit all. We often need alternative technologies, 
sometimes simple vs advanced ones, or cut up the problem space in 
different ways: different stepping stones, but leading to the same 
goal. So we have CSS and XSL, XSLT and XBL, SMIL and TT-DFXP, XML and 
EXI, SVG and InkML. They overlap, but they don't make each other 
redundant.



Bert

PS. Sorry for all the philosophy. There is so much e-mail on the 
www-style list, that I need to relax my mind a bit, before I have the 
courage to read another long thread. 400 messages in two weeks, and in 
the middle of the holiday period for most of us. That's not a sign of 
popularity anymore, it's a DoS attack...

-- 
  Bert Bos                                ( W 3 C ) http://www.w3.org/
  http://www.w3.org/people/bos                               W3C/ERCIM
  bert@w3.org                             2004 Rt des Lucioles / BP 93
  +33 (0)4 92 38 76 92            06902 Sophia Antipolis Cedex, France

Received on Thursday, 17 July 2008 18:27:20 UTC