Re: Tim Berners-Lee on the Semantic Web

From: Jonathan Rochkind <rochkind_at_nyob>
Date: Tue, 27 Oct 2009 12:24:06 -0400
To: NGC4LIB_at_LISTSERV.ND.EDU
That's not an unreasonable thing to do, but it's also not necessary.  
GUID's are neat becuase they guarantee that they are globally unique, 
they won't conflict/collide with a GUID assigned by someone else.

But when you put it in a URI, you don't need a guarantee of global 
uniquness. Becuase the URI already has that built into it -- you can't 
(or at least you violate a lot of ground rules if you) mint a URI 
beginning with 'http://clasate.cimec.ro/' unless you own the 
clasate.cimec.ro domain.

So there's no need for http://clasate.cimec.co/Detailu?[guid], because 
as long as you make sure what goes after the question mark is unique 
within _your_ system you guys at clasate.cimec.co are using, you're 
good, you don't need a GUID. But, hey, if it's convenient for you to use 
one, there's nothing wrong with it.

[ There is potentially something wrong with having your URI begin 
http://clasate.cimec.co/Detailu.asp though.  Some day you might stop 
using asp. You don't want all your URI's to be ruined when you do so. 
Much better to keep your URI's technology independent.  
http://clasate.cimec.co/Detalu?id is much better.  Although the REST 
fanatics will tell you it should be http://clasate.cimec.co/Detailu/id 
instead.  (Depending on what 'detailu' means, I don't know).  I think 
it's arguable how necessary it is, but these could be considered 
analagous to 'level of normalization'.   First just get the .asp out of 
there, we know that's bad. Then we can argue about how neccesary it is 
to try and make it REST-like.]

Incidentally, this is one of the things I really LIKE about URIs.  I 
don't actually care about the resolvability as much as, say, rsinger, 
I'm only slightly buzzed on the linked data juice, not fully sloshed. 
But if I just have an OCLC number, say "23233455", it's very easy for me 
to lose the context that that IS an OCLC number.   I need to make sure 
not to confuse it with the LCCN "23233455". If instead I'm using 
"http://www.worldcat.org/23233455" as an identifier, the context comes 
with it, it's not gonna get seperated, and it's not gonna collide with 
"http://id.loc.gov/lccn/23233455".  It's globally unique _without_ 
needing to use something cumbersome like a GUID.

Jonathan


Dan Matei wrote:
>> -----Original Message-----
>> From: Next generation catalogs for libraries 
>> [mailto:NGC4LIB_at_LISTSERV.ND.EDU] On Behalf Of Ross Singer
>> Sent: 27 octombrie 2009 15:55
>>
>>     
>>> What about GUID [Global Unique Identifier] aka UUID ?
>>>
>>> E.g. 73f6abe5-ce82-4f92-88cc-59e81b858664
>>>       
>> I would much, much rather have a URI that could potentially 
>> give me resource description back.
>>     
>
> Me too ! But GUIDs could be used in URIs.
>
> E.g.
>
> a) on our site:
> http://clasate.cimec.ro/Detaliu.asp?k=5C427C0714904A60B5866547C861EDA6
>
> b) the same Bruegel in Europeana:
> http://r.culturalia.ro/VizualizareImagine.aspx?guid=5C427C07-1490-4A60-B586-6547C861EDA6
>
> Dan
>
>   
Received on Tue Oct 27 2009 - 12:27:58 EDT