Use of LSIDs in RDF

Hello, 

I'm new to this list so forgive me if this is not the right forum to
discuss these issues, nonetheless...

The Life Science Identifier (LIDS) proposal [1] addresses the need for
persistent and  unique identifiers  for biological resources.  While a
solution for this is needed, I am wondering whether the use of URNs as
defined by the LSID spec is appropriate for the semantic web ?

For example, I have some expectation that URIs in RDF can be dereferenced
to provide either a description of the resource or the resource itself
from the Internet [2]. At present HTTP URIs are the best game in town for
this. So why is LSID proposing URNs as a solution ? 

The main reason against using HTTP URIs is persistence and the separation
of a name and network location [3], according to the spec, but maintaining
persistence is largely an organizational (social issue) [4]. 

But for the moment let's just say that URNs are more persistent. The LSID
spec provides a resolution scheme, here is an outline:

1. HTTP URIs are not persistent

2. Invent a new naming scheme LSID

3. Invent new infrastructure for dereferencing LSID (DDDS)

4. Deploy new infrastructure  

5. Dereferencing a LSID results in a HTTP URI 

Which brings me to my question, why not just use persistent HTTP URIs
in the first place ?

_greg

[1] http://www.omg.org/cgi-bin/doc?lifesci/2003-01-06
[2] http://www.w3.org/2001/tag/webarch/#identification
[3] http://gbiv.com/protocols/uri/rev-2002/rfc2396bis.html#URLvsURN
[4] http://www.w3.org/2001/tag/webarch/#representation-management

-- 
Greg Tyrelle