Section 1.2.2 (Re: draft-fielding-uri-rfc2396bis-03)

> Please note that all issues have been fixed or closed.  If you'd
> like to raise a new issue or reopen an old one, please do so
> within the next two weeks.

From a Semantic Web perspective, I'm fairly happy with section 1.2.2
("Separating Identification from Interaction"), but there are still
a two bits of wording I'd like to see changed.

1.  The use of "denote" in the second paragraph seems pretty far off.
    Yes, we use URIs to denote resources, but that's nothing like
    performing an operation on a resource.  Am I performing an
    operation on my dog when I tell you he's shedding a lot today?  I
    don't think so.  So can we strike "denote"?

    In most straight semantic web applications (not semantic web
    services apps), the only operation likely to occur is retreival,
    as described in the third paragraph of 1.2.2.  If you want some
    text suggestive of RDF, then "describe" could be added to the
    list of operations.

2.  I agree with the premis of paragraph four, that URIs are meant to
    be "late binding", identifying into the future, but I disagree
    with the conclusion that "the resource referred to by the URI is
    actually a sameness of characteristics as observed over time".
    Maybe that phrase can just be dropped?   

    My view is that most traditional http URIs denote sources of
    information, as described in paragraph three of 1.2.2.  If a
    source is consistent in some regard over time (always describes
    the weather in LA, etc) then it makes sense to use its URI; if
    it's not, then it doesn't make sense to use its URI.  So if you
    see a URI actually used to name a source of information it
    probably means the URI denotes a consistent-over-time one.

    IMHO, information sources can be well-characterized as kinds of
    services (in the sense used in Web Services, loosely "things which
    do things for clients", which might be more gramatically called a
    "service providers"); specifically they are information-providing
    services.  This characterization suggests that many URIs (perhaps
    all "working" ones?) denote services.  But here we're clearly out
    of the realm of consensus and out of what needs to be covered in
    this RFC.

    My point is: please don't put text in the URI specification which
    suggests that URIs denote "samenesses" when there are plenty of
    other good and different ideas about what URIs denote.

Thanks for listening.  Please let me know if you are not persuaded and
plan to keep the text I suggest dropping.

    -- sandro

Received on Saturday, 14 June 2003 16:47:21 UTC