Re: identifier to use

Phillip Lord wrote:
> Actually, LSIDs are domain specific, or rather they were designed to support
> the needs of the Life Sciences; this is not to say that different domains do
> not have the same needs. 

You're right, that's a better way to put it!


> Look at DOIs and LSIDs. They are different, they emphasise different
> things. LSIDs are based around a set of objects which potentially might be
> very large and which might exist in many versions. So LSIDs have two-step
> multi-protocol resolution. They have version numbers integrated. They exist
> in a world where services disappear. So LSIDs have a fail over mechanism. 

If I have an LSID like urn:lsid:uniprot.org:uniprot:P12345 and uniprot.org 
disappears (assuming there was even a resolver running there in the first 
place), how is that URI going to be more useful than a simple HTTP URL like 
http://purl.uniprot.org/uniprot/P12345? If you are lucky and happen to have 
a copy on some local server (which you may prefer to use even otherwise), 
what's the big deal with using a normal, off-the shelf URL rewriting proxy?


> DOIs worked because there are actually relatively few publishers and they
> moved on mass. In biology there are many more service providers, and most will
> not adopt something till it looks stable and until people really bitch about
> wanting it. 

Perhaps as important as "what are the requirements" is "what is practical".

Received on Wednesday, 22 August 2007 17:38:02 UTC