RE: The self-describing web...

That should be true and probably is.  One does have to 
consider the media, conditions for release, stability of 
agreement, retrievability, etc.  The release of the 
information about the condition of the miners in West 
Virginia is revealing.  A web app can be quickly changed. 
It's distribution is in the 'click market'.  A newspaper 
can't.  So assertions have the property of stability over 
time and distribution space.   Bindings of namespaces 
have similar properties given versions and contexts. 

One may wish to consider ranges of 'strong' and 'weak' properties.

http://cogweb.ucla.edu/CogSci/Talmy.html

Given the principle of least power, would it be architecturally 
correct to say one favors weak properties ("strict in what is sent, 
liberal in what is accepted")?

I wonder if given a read/write web with imperatives, the 
notion that the receiver always has autonomy to interpret 
is flawed; that is, in effect, you are implicitly stating a 
relationship exists between sender and receiver which is 
the interpretive context ("an author whom you trust").  
If that relationship is weakly established, the range of 
interpretation is as well.  Conditions for setting the 
relationship and its strength are useful (eg, inside 
the firewall, external to firewall, obligated by 
contract (implicit or explicit)).  Ummm.. yuck... I 
feel the lawyerly personna emerging around topics of 
establishment of duty.

len


From: www-tag-request@w3.org [mailto:www-tag-request@w3.org]On Behalf Of
Norman Walsh

Ah. Sorry. That's not what I meant. I wasn't trying to say that
malicious behavior doesn't exist, it surely does. What I meant to say
was more the other way around: if an author whom you trust is not
malicious publishes a page and you are not reading that page for some
specific purpose known a priori to yourself (such as counting words or
building a search index or checking to see if the content is written
in French), then you ought to be able to follow your nose and figure
out what the content says. And what you determine it says ought not
usually to be too far off the mark of what the author actually thought
he was saying.

Received on Wednesday, 4 January 2006 18:53:20 UTC