W3C home > Mailing lists > Public > www-tag@w3.org > January 2008

Re: Resource-Type Revisited (httpRange-14)

From: Sean B. Palmer <sean@miscoranda.com>
Date: Fri, 4 Jan 2008 19:07:41 +0000
Message-ID: <b6bb4d890801041107n55dc5bf2n1891e387e4fef695@mail.gmail.com>
To: "Richard Cyganiak" <richard@cyganiak.de>
Cc: www-tag@w3.org
On Jan 4, 2008 5:40 PM, Richard Cyganiak <richard@cyganiak.de> wrote:

> the 303 style makes this hard. This is a very good point, and
> one that has not occurred to me before.

Yeah, nor me until last month. Suddenly I was displaying this list of
potential sources of RDF, and there was no way for anyone to give them
titles or keywords. It'd be like if we couldn't use link texts in <a>
elements in HTML!

> All of this is a non-issue when hash URIs are used.

Yep.

> This raises the question: Why pile hack upon hack to make
> slash URIs work as semwebby identifiers?

Very good question. I'd just like to note as a clarification first
that my proposal for a :from property, so that we can easily give
redirected IRs properties in RDF, is of course independent of
Resource-Type. We need that no matter what, since we already have the
303 technique.

I think there are at least two reasonable arguments why we ought to
use slash URIs, motivating the need for hack-upon-hack. The first, and
weaker, is that they're simply natural. Dublin Core was the first
ontology to use slash URIs, I think. I've no idea why they did, but I
just assume they figured that HTTP URIs are the words of the web, so
why not use them?

Once that mentality had started to take hold, it was irreversible. RSS
1.0 used slash URIs, and then FOAF—and this was all before 2002. Slash
URIs had to be accounted for, and we ended up with httpRange-14.

The second, and stronger, argument is that they are in some cases
better to use. The idea is that if you have a namespace which has a
million components in it, it's impractical to serve one big file there
which contains info about all of the components. Even if it's just
loads of rdfs:seeAlso arcs, it'll be a big file.

The counter-argument is that you can break the files up into chunks.
But then you end up having loads and loads of namespaces, and really
you should be able to get away with one. And you can get away with one
if you use a slash URI.

Ironically, for Dublin Core, RSS, and FOAF none of this was necessary.
They *should* have used hash namespaces because there are only a few
dozen components in their namespaces. Oh well. But the syntactic
convenience of namespaces is a fairly compelling argument to me... not
massive, I must admit, but I'm being a descriptivist here. It's used,
and people will continue to follow the example of DC, RSS, FOAF et al.

-- 
Sean B. Palmer, http://inamidst.com/sbp/
Received on Friday, 4 January 2008 19:07:55 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:55 UTC