W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > August 2007

Re: identifier to use

From: Hilmar Lapp <hlapp@duke.edu>
Date: Thu, 23 Aug 2007 10:35:40 -0400
Message-Id: <5E3E0E98-8985-46A2-9E3F-8C5A0A5B1C36@duke.edu>
Cc: Phillip Lord <phillip.lord@newcastle.ac.uk>, public-semweb-lifesci hcls <public-semweb-lifesci@w3.org>
To: Eric Jain <Eric.Jain@isb-sib.ch>

On Aug 23, 2007, at 9:58 AM, Eric Jain wrote:

>> Digital archives will use opaque identifier systems that aren't  
>> HTTP URIs whether the W3C likes it or not - they look at time  
>> horizons beyond our lifetimes, when HTTP may not even exist  
>> anymore. The need for GUIDs existed before HTTP URIs and will  
>> continue to exist afterwards.
> For closed archives I guess it's easier, as you can prescribe  
> whatever identifier system you like, no matter how exotic.

I'm not sure what you mean by "closed archive."

> But if you are going to archive public data from many many sources,  
> I imagine you want to go with the greatest common denominator,  
> which is HTTP URIs for the Semantic Web.

Almost no database, data center, or publisher uses HTTP URIs for  
identifying their digital objects, and stable HTTP URIs at present  
aren't adopted or the common denominator for identifying digital  
objects in the life science domain either. And that's not b/c it is  
technically difficult to do so, or because nobody knew that stable  
URIs are desirable.

I think that's where the requirements of digital archives and those  
of the semantic web probably deviate. 404s are rampant (and not  
because nobody would care anyway) but the web is still useful. Using  
HTTP URIs for the semantic web will result in the same abundance of  
resources that fail to resolve although they could, and maybe that's  
just fine. But for digital archives it's not.

On Aug 23, 2007, at 9:22 AM, Xiaoshu Wang wrote:
> [...] I think a more reasonable approach is (1) we should allow the  
> invention of name but (2) discourage the invention of  
> transportation protocol.  Any newly invented URI scheme should be  
> required to specify a straight forward mapping to HTTP  
> transportation protocol.

This makes a lot of sense. I think it will need to be understood  
though that any such mapping may be temporary (on the scale of  
years), but that's probably not an issue at all.

: Hilmar Lapp  -:-  Durham, NC  -:- hlapp at duke dot edu :
Received on Thursday, 23 August 2007 14:36:10 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:20:29 UTC