Paul Johnson wrote:
On Fri, May 26, 2006 at 05:25:49PM -0400, A. Rick Anderson wrote:
Gordon Clarke wrote:
Isn't the Open Source community great.
Thanks Paul for another example of generating guids!!!
This algorithm doesn't seem very close to the DCE/IETF/OSF standard:
http://en.wikipedia.org/wiki/UUID.
Once you pay the price for 128 bit key, it seems rather silly not to use
a standard that just about always will guarantee uniqueness.
I may be off base here. I wrote a Java implementation of the ISO
standard several years ago, and at that time, the standard didn't
support multiple algorithms.
You obviously know more about this that I do, since I am merely using
the Data::UUID implementation on CPAN, but as I understand it, this
implementation is based on
http://hegel.ittc.ku.edu/topics/internet/internet-drafts/draft-l/draft-leach-uuids-guids-01.txt
which seems to be an early version of http://www.ietf.org/rfc/rfc4122.txt
which is referenced on the wikipedia page you mentioned.
Now the first document specifies a UUID "is either guaranteed to be
different from all other UUIDs/GUIDs generated until 3400 A.D. or
extremely likely to be different (depending on the mechanism chosen)".
The second specifies that the UUID "can guarantee uniqueness across
space and time". Since the latter claim is obviously impossible in a
finite number of bits, (and since I've not investigated the matter in
detail), I'm unsure whether the algorithm has changed or whether some
marketdroid got to the rfc between these documents. But a quick scan of
the rfc does seem to allow for the possibility of multiple algorithms
anyway.
Leach has been pushing multiple algorithms since the mid-nineties, or
earlier. But at the time when I was implementing it, I was at IBM and I
was reading the actual ISO/DCE specification. At that time, there was
just the node-based algorithm. Since Java can't get the MAC address, I
had to settle for using the IP address and rely on the clock sequence
(based on a crypto-strength random number) to guarantee uniqueness.
The actual ISO specification was much easier to read that what Leach was
pushing at the time. Because of Leach's fixation on multiple
algorithms, his write-ups are always way more complicated and convoluted.
--
A. Rick Anderson
_______________________________________________
Ldsoss mailing list
[email protected]
http://lists.ldsoss.org/mailman/listinfo/ldsoss
