SourceForge : Post

Project: Editor Discussion > INFO: OGSA: Glosary of Terms > name/id/address definition > List of Posts

Forum Topic - name/id/address definition: (2 Items)

View: as

None

09/07/2004 4:41 PM

post4471

name/id/address definition

I like the definitions. For name/id/address, I am missing some definition
of uniqueness and scope. Does OGSA assume globally unique ID\'s?
If so, it should be statet. Prior GT\'s had those, also OGSI.
This should be statet because it has significant implications.
I would rephrase: \"Name is an attribute used to uniquely identify an
entity within a {local/global/definable - TBD.} scope.\".
\\\\
Names are typically only unique within a certain scope/namespace.
Seems that OGSA doesn\'t have an explicit notion of namespaces
(I would ask, why)? If there is no notion of namespaces, how conflicts
are resolved when, for instance, 2 people pick the same name for machines?
\\\\
The term \"abstract name\" seems to be what is typically called an Id. I would use that familiar term rather than 
abstract name.
\\\\
Address = location, that\'s correct. 

Sven Graupner
 -Hewlett-Packard Laboratories
 -Internet and Computing Platforms Research Center (ICPR)
 -1501 Page Mill Road, MS 1182, Palo Alto, CA 94304, USA
 -sven.graupner@hp.com -- phone: 650.857.3849 -- fax: 650.236.1612

Jem Treadwell

01/25/2005 2:45 PM

post4472

Response to Sven Graupner's comments

Hi Sven, thanks for your comments.  Andrew Grimshaw has been leading the OGSA-WG's work in the naming area, and we did 
make some changes in the final version.  Below are Andrew's responses to your comments.  Andrew's remarks are marked 
with <<ASG: ... >>.

- Jem


For name/id/address, I am missing some definition of uniqueness and scope. Does OGSA assume globally unique ID's?  If so
, it should be stated.  Prior GT's had those, also OGSI. This should be stated because it has significant implications. 

<<ASG: In discussions we have said that abstract names ARE globally unique. It is not technically difficult. >>
I would rephrase: "Name is an attribute used to uniquely identify an entity within a {local/global/definable - TBD.} 
scope."
 
Names are typically only unique within a certain scope/namespace. Seems that OGSA doesn't have an explicit notion of 
namespaces (I would ask, why)? If there is no notion of namespaces, how conflicts are resolved when, for instance, 2 
people pick the same name for machines? 

<< ASG: At this level  they can't, they're globally unique.>>

The term "abstract name" seems to be what is typically called an Id. I would use that familiar term rather than abstract
 name.

<< ASG: Identity was felt to be a term tied in with crypto  and overloaded. So we chose abstract name.>>

Return