Re: Syntax and semantics

One of the odder bits of fallout I'm hearing in response to the insistance
that everything which looks like a URIRef must have URIRef semantics,  is
that developers who were planning to use URIs in their designs have started
talking about switching to other notations in order to dodge this bullet.

I've heard several folks  propose cutting over to Java package names. Or
perhaps URLs, if they don't carry the same baggage. Or some brand-new
notation which could easily be converted to a URI when that's desirable,
but which would not _be_ a URI and thus wouldn't get entangled in a similar
argument.

I freely admit that I don't have enough data points to claim that this is a
trend. But it makes me very nervous. If the attempt to defend
syntax-implies-semantics drives people to custom syntax, that might well do
more harm to the web metaphor than allowing syntax and semantics to be
applied separately would.

______________________________________
Joe Kesselman  / IBM Research

Received on Wednesday, 17 May 2000 15:08:28 UTC