W3C home > Mailing lists > Public > uri@w3.org > July 2009

Re: URI for abstract concepts (domain, host, origin, site, etc.)

From: Xiaoshu Wang <wangxiao@musc.edu>
Date: Wed, 01 Jul 2009 20:54:55 -0400
Message-ID: <4A4C055F.7070502@musc.edu>
To: Erik Wilde <dret@berkeley.edu>
CC: "'URI'" <uri@w3.org>
Erik Wilde wrote:
> hello.
>
> Xiaoshu Wang wrote:
>   
>> In other words, the URN is an HTTP-URL sans "http:"
>> There are several advantages of this.  Although an HTTP-URL can be taken 
>> as a URN as well as a URI.  But this dual mode makes people very 
>> uncomfortable.  This scheme-less URN helps easy the issue.
>>     
>
> i am nor sure i understand this. URNs were intended to denote things 
> that have no clearly defined way of retrieval, but the notion that 
> locators and names are two entirely different things never was really 
> true, which is probably the reason why the whole URN thing did not 
> happen. but by using domain names in what you call "URNs", you put quite 
> a bit of internet-specificity into what should be an abstract naming 
> scheme. how would you, for example, in that approach map ISBN numbers to 
> your proposed URN structure?
>   
First, I think that an absolute sense of URN is of no use.  People must 
associate a name with some kind of access-protocol,  walking, running, 
driving, seeing etc. to interact with the desired referent in order to 
know the referent.  My URN is in the sense of it is independent of a 
protocol, but does not mean that there is no protocol to support it.

Second, the relation between a name and its transportation protocol must 
be maintained/defined somewhere, otherwise a name is useless.  What I am 
suggesting is that to store the relationship between a URN and its URL 
in humans as opposed to be defined as a technical specification.  
Because one we do the latter, we makes URN into a URL. To put it in the 
hands of human knowledge makes the binding loose enough.  I think given 
a //w3.org nowadays, most people will try to access it with "http://".  
Most companies now days that have omitted the "http://" altogether and I 
doubt that has hurt their chances of making people go to their web site.  

Of course, URN is still dependent on domain name, which is internet 
specific.  This specificity, in my opinion, is a necessity because it 
grounds the Web onto the internet. If you insist that a name must not be 
specific to any technology, then I cannot argue any other way around.  
But that will put us back to the old days.  However, if you think that 
the internet will be there for quite a while, then I don't think this 
internet specificity hurts more than it helps. 

For your ISBN example, I don't know who actually is responsible for 
issuing that.  But I doubt it will be difficult to persuade them 
nowadays to host a web site.  Let's say that they own the domain of 
"isbn.org", then they can use that as the default namespace for all ISBN 
numbers, such as an URN like isbn.org/12345678 would denote a specific 
ISBN number. 

Xiaoshu
Received on Thursday, 2 July 2009 00:55:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:42 GMT