RE: The Semantic Debate

Jonas Sicking wrote:
> 
> Not sure what you mean. You gave 3 examples right below, "sarcasm",
> "joke" and "critical". I believe "ship" has also been mentioned on
> this list.

You are missing the point.  How do we now translate those <markups> to
machine readable value?  Do you expect the average person to view the
source-code to see if I am being <sarcastic> or if the issue is <critical>?

> 
> Unfortunately we don't have the luxury of having humans interpret the
> tags we come up with, so I think we have to limit ourselves a bit.

Perhaps, although RDF removes that limitation to some extent.  That a
"common collection" of reserved concepts exist is fine and good, and I would
support that idea (although those concepts could then be used with either
@role or @class, so long as it's standardized)

> 
>> Or are you suggesting adding both?  That to me seems
>> counter-productive and more confusing.
> 
> Why? That way people that find RDFa is too complex can limit
> themselves to the predefined (and prefixed) classes that we put in
> the spec. For the people that want the full power of RDFa can use the
> role attribute. 

Jonas, I'm sorry if we are simply not understanding each other, but this
makes no sense.  There is the means to identify semantic value, and then
there is the value itself.  If you desire a collection of common meanings,
then I for one would have no issue with that - heck, I think a lot of the
accessibility advocates would volunteer to help create that collection.

The issue becomes the vehicle that delivers those values: the already
used/abused @class attribute, or the new, still fresh and tender @role
attribute.  Read the spec for @role, and you will see that it already uses
Qnames and an XML namespace and RDF and all that jazz - whoa, I know, scary
for the average web punter who just wants to make web pages.  But that
doesn't matter, because with the roles defined, they can just simply go
ahead and use the pre-defined collection: they don't have to think about the
RDF stuff.  

There has been a fair bit of talk about WYSIWYG editors in this debate, so
the backend code that gets spit out is secondary to the ability for the
author to "just do it".  The next-gen WYSIWYG editing tool has a "Meaning"
button that allows you to choose from the common collection - how is this
hard?

JF

Received on Monday, 7 May 2007 19:19:55 UTC