legalxml-courtfiling — archive

RE: [legalxml-courtfiling] Globally unique identifiers - approach

From: James Cabral
To: [email protected]
Date: Thu, 12 May 2005 22:31:00 -0000

MHonArc v2.5.0b2 -->

legalxml-courtfiling message

Subject: RE: [legalxml-courtfiling] Globally unique identifiers - approach

From: "Cabral, James E." <[email protected]>
To: "Cusick, James" <[email protected]>,<[email protected]>
Date: Thu, 12 May 2005 15:30:28 -0700

James,

I did not mean to suggest that the proposed approach would not work. I just wondered whether there was any value in having the timestamps for things that are static. I was unclear, after my first reading of the RFC, whether any components of the UUID, such as the timestamp, were optional. After re-reading it, it appears that the time timestamp is a required component.

jim

From: Cusick, James [mailto:[email protected]]
Sent: Thursday, May 12, 2005 2:42 PM
To: Cabral, James E.; [email protected]
Cc: John Greacen
Subject: RE: [legalxml-courtfiling] Globally unique identifiers - approach

Jim, I would propose that this approach will work for all artifacts or objects no matter when or how often they are created or accessed. Can you give me a scenario in which this approach does not work? For things that do not change often the ID will still be unique even though the timestamp is from some distant past moment. Did I miss your point? regards, james

From: Cabral, James E. [mailto:[email protected]]
Sent: Wednesday, May 11, 2005 8:53 PM
To: Cusick, James; [email protected]
Cc: John Greacen
Subject: RE: [legalxml-courtfiling] Globally unique identifiers - approach

James,

That approach seems appropriate for identifying things that occur at a specific time, e.g. Filing Identifiers, but do we need a timestamp for things that do not change often (e.g. FilingAssembly or FilingReview ID)?

jim

From: Cusick, James [mailto:[email protected]]
Sent: Wednesday, May 11, 2005 12:41 PM
To: [email protected]
Cc: John Greacen
Subject: [legalxml-courtfiling] Globally unique identifiers - approach

John and all,

At the New Orleans meeting I agreed to propose an approach to generating globally unique identifiers. After searching around on this topic I would like to recommend using UUIDs as per the IETF DRAFT RFC. This approach requires NO global coordination.

Please see:

http://www.ietf.org/internet-drafts/draft-mealling-uuid-urn-05.txt

This RFC is currently in draft status and is based on Unique ID generation technique used in Open Software Foundation's Distributed Computing Environment. The RFC guarantees that ID's generated using the recommended algorithms will be unique across space and time.

This RFC specifies 4 algorithms to generate universally unique ID's. The following subset is recommended for LegalXML standards:

1. System timestamp up to a 100 nanoseconds

2. MAC address of the machine where UUID is generated

3. Use of Random or pseudo random numbers

Finally, the RFC contains C code sample on how to generate UUID's based on the generation algorithms listed above.

Please let me know what you think and how you would like to see this phrased within Blue.

Regards,

james