[Dev] Re: large random IDs (4suite uuid.py)

[Uche Ogbuji]

> John Anderson had expressed a desire for IDs bigger than 128 bits.
> he might be right, if Chandler is deployed very widely and there are huge
> numbers of IDs that might be visible from a single client context.  We
> would prefer chance of collision is very close to zero. if we could
> guarantee very random numbers, than 128 bits is very big.  But we might
> want 192 bits to be on the safe side.

I forgot the scenario that is sometimes trotted out to capture the 
unlikeliness of collisions in UUIDs generated frm a truly random engine.  It's 
something like, if all the atoms in the galaxy were all generating a UUID per 
second, it would still be years before a collision.

IOW, I *really* don't think you need to worry about collisions, even with the 
Leach/Salz node algorithm, never mind the random version.

I regret that I am short of time today to rigorously look these things up 
again, but I did research them a decent amount when I first wrote Uuid.py.


-- 
Uche Ogbuji                                    Fourthought, Inc.
http://uche.ogbuji.net    http://4Suite.org    http://fourthought.com
Introducing N-Triples - http://www-106.ibm.com/developerworks/xml/library/x-thi
nk17/index.html
Use internal references in XML vocabularies - http://www-106.ibm.com/developerw
orks/xml/library/x-tipvocab.html
EXSLT by example - http://www-106.ibm.com/developerworks/library/x-exslt.html
The worry about program wizards - http://www.adtmag.com/article.asp?id=7238
Use rdf:about and rdf:ID effectively in RDF/XML - http://www-106.ibm.com/develo
perworks/xml/library/x-tiprdfai.html
Keep context straight in XSLT - http://www-106.ibm.com/developerworks/xml/libra
ry/x-tipcurrent.html
Using SAX for Proper XML Output - http://www.xml.com/pub/a/2003/03/12/py-xml.ht
ml