W3C home > Mailing lists > Public > uri@w3.org > December 2007

Re: URIs & Namespaces

From: Erik Wilde <dret@berkeley.edu>
Date: Sun, 09 Dec 2007 23:02:47 -0800
Message-ID: <475CE497.6020506@berkeley.edu>
To: uri@w3.org
CC: Mike Schinkel <mikeschinkel@gmail.com>

hi mike.

> Currently there already exists a universal way of defining vocabularies for
> placenames; it is called Wikipedia.  Can you give me an example of a named
> place that does not have a URL on Wikipedia?

sure, just tell me when to stop:

- the pool where i swim over lunch
- the korean place where i like to eat after swimming
- almost all neighborhood names in the bay area

and then of course all the "named places" where the place is more of a 
social concept than the landmark kind of place that gets into wikipedia. 
wikipedia is great, but if you could listen how people use place names 
in their everyday conversations, most of the place names they use you 
wouldn't find in wikipedia.

> Namespaces solve some problems in many contexts, and create huge problems in
> others. What I envision would actually use namespaces, but if namespaces
> were allowed to be infinite, we'd have the same problems we have with
> namespaces in XML; it's almost impossible to associate between two
> namespaces without a lots of human effort involved each time.

yes, but that's the effort that comunications always require: you need 
to understand the concepts referred to by your peer, otherwise you are 
not going to understand them. namespaces in real life are infinite 
(almost), if you look at all the languages and cultures and social 
groups. and sometimes they make conscious efforts to invent their own 
"namespace", such as young people talking in a way which is hard to 
understand for adults. i think the important thing is to have some way 
to be able to recognize when i could understand you, and that can only 
happen when i can recognize the vocabulary you are using.

> What you are pursuing and want I am pursing have different goals. What I
> envision can accommodate and even enable what you want, but it sounds like
> you may have an ideology that disagrees with my approach.  If so, let's just
> agree to disagree as debates between ideologies are rarely productive.
> However if you do see my ideas as having potential merit, or if someone else
> sees them as having potential merit, I'd be very interested in discussing
> further, though quite possibly off list.

i wouldn't call my idea an "ideology", and on this mailing list i am 
simply looking for an answer to the rather technical question:

assumed it makes sense in some space of identifiable resources to 
partition these into namespaces, identified by names within that 
namespace. and i don't want to maintain a registry for namespaces. what 
would be the most uri-like way to do it? or is that something that out 
of general principle never should be done within any uri scheme?

cheers,

dret.
Received on Monday, 10 December 2007 07:03:07 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:11 UTC