Re: Entered lots of text into draft

On Tue, 2009-10-27 at 13:22 -0400, Jonathan Rees wrote: 
> http://www.w3.org/2001/tag/awwsw/http-semantics-report.html

1. Nice start on the draft!  One suggestion: I think the intro needs a
little more context setting for readers who are not familiar with this
work.

2. I've been trying to conceptualize what I think we (as a group) should
deliver.  The most important thing IMO is to propose an ontology and
rules that indicate (in RDF) what assertions an HTTP server is making
when it sends a client various HTTP responses, such as a 200 response, a
301 response, a 303 response, etc.

Although I think a broader ontology that attempts to relate existing
notions of "information resource" (from AWWW), "network data
object" (from RFC2616), etc., would be somewhat interesting and may be
helpful as a guide to people who ponder the definitions of such terms, I
do not think such an ontology is ultimately what's needed nor what we
should focus on.  Here's why:  

- Various documents related to web architecture define notions that are
semantically overlapping.  For example, AWWW uses the term "information
resource",[@@] RFC2616 uses the term "network data object"[@@] and
Fielding uses the term "resource"[@@] in overlapping ways.

- For many purposes, these terms can be considered synonyms even though
their definitions are not the same.  For example, if we restrict our
focus to HTTP interactions, then x is an information resource in the
AWWW sense if and only if it is a network data object in the RFC2616
sense and if and only if it is a resource in the Fielding sense.  

- However, there are edge cases for which an application may wish to
distinguish between such definitions.  For example, even though x is a
network data object in the RFC2616 sense, x is not a resource in the
Fielding sense.

- Sameness within an application domain does not require identical
definitions.  It is normal that some applications will treat certain
terms as synonyms while other applications treat those terms as
different.  owl:sameAs and the like can indicate when two terms are
being treated as synonyms.  

- Therefore, the "problem" of disparate definitions between these
various documents is not really fundamental problem, though it may be a
bit annoying.  It is just life as usual in the RDF world: terms that one
application treats as synonyms may be treated as different by another
application.  It is normal and natural for different application domains
to define similar notions differently, because they have different
concerns.  Ultimately, this is no different from one application domain
defining a class "person" one way, while another application defines a
class "person" in a completely different way.  Some applications may
treat these two as the same, and others may treat them as different.

In short, I think our most important deliverable should be to propose an
ontology and rules that indicate (in RDF) what assertions an HTTP server
is making when it sends various HTTP responses.  That would be a
tangible contribution.   And I think the rules that I drafted some time
ago
http://esw.w3.org/topic/AwwswDboothsRules
and the shell script that you drafted, would be good starting points.


-- 
David Booth, Ph.D.
Cleveland Clinic (contractor)

Opinions expressed herein are those of the author and do not necessarily
reflect those of Cleveland Clinic.

Received on Tuesday, 17 November 2009 04:00:54 UTC