"Publisher" -> "Host" and "Namer" (Bearing on UseImpliesConsent)

Bernard Vatant wrote:
> There are basically two stakeholders in the use of a URI. Let me call them
> hereafter "Publisher" and "User".

I think we need to separate "Publisher" into "Host" and "Namer", each
of which has a different but potentially privileged position for
saying things about what a URI means.  The namer is the entity which
first used some URI as a name for something.  The host is the entity
which answers web queries using that URI.  It seems pretty clear to me
that the namer has the moral high ground in saying what a URI means,
while the host has the big guns.

When Eric Miller names himself "http://www.w3.org/People/EM/contact#me",
the namer is certainly Eric himself, and so is the host.  But if
w3.org changed hands tomorrow, Eric would no longer be the host.  If
the W3C management decided that Eric didn't get to control
/People/EM any more, then Eric would no longer be the host.  If
someone broke in and altered the content without Eric's permission, Eric
would have stopped being the host and might not even realize it.
Really, the host is a distributed computer system, but when we talk
about what it "says", we know it's just acting as an agent for something
else, saying what it's been told to say, so lets focus on the
identify of whatever told it to say that.

There is some cloudiness around "first use" in the definition of
namer, since the first use might well be in private.  If the first use
is big, public, on record, and clearly constrains the meaning, the
namer's moral high ground is a lot higher.

There is potential for the namer and the host to say things which are
different from what they each meant to say.  I'm not sure how to think
about that.

Am I right that the first use of a name is always right?  Perhaps not
a use of a mailto: URI to denote an integer....

As for the host, I don't think the host has much "rightness", just a
lot of marketing muscle.   Might makes right?  :-)

Fortunately, in practice, the namer and host will often be well
aligned.  The possibility of disagreement has been fueling a lot of
the debate here, I think: the intuition that the namer is right by
defintion has been clashing with the fear that the host could wind up
saying bad things.

Operationally, I might codify this as: one MUST NOT say things using a
URI term which are known to be logically inconsistent with its first
use (its introduction); one SHOULD NOT say things which are logically
inconsistent with how the host uses the term.  (I wonder if saying you
shouldn't be logically inconsistent with some content is more
palatable than saying you must assert that content.  I suspect it is
more palatable, and nearly as useful.)

    -- sandro

(Yeah, I'm trying to get my ideas out now, because I'll be mostly
focussed on process when chairing the meeting tomorrow.)

Received on Thursday, 25 September 2003 15:32:39 UTC