On Sunday 07 September 2008 09:25:18 Hari krishna Anandhan wrote:

> Any thoughts on scenarios where the above would not work?

Hi,

just creating a UUID as URIs (nepomuk:/<UUID> ) should be enough in theory. 
They are invented for creating global unique IDs for creating ids decentral 
and later put these data repos together without any collision.

>From Wikipedia:

"In other words, only after generating 1 billion UUIDs every second for the 
next 100 years, the probability of creating just one duplicate would be about 
50%. The probability of one duplicate would be about 50% if every person on 
earth owns 600 million UUIDs.

The odds of a properly generated duplicate UUID being used in error in any 
practical circumstance is so low that any other work to improve software 
reliability would clearly be a much higher priority."

http://en.wikipedia.org/wiki/Universally_Unique_Identifier


The question is, what UUID algo to use.  The wikipedia article says KDE is 
already using one...

btw: The one from Qt is "broken" I tried to use it, but got collisions even on 
the same computer after a few thound created ones. 

>From the Qt docu:

"The UUIDs generated are based on the platform specific pseudo-random 
generator, which is usually not a cryptographic-quality random number 
generator. Therefore, a UUID is not guaranteed to be unique cross application 
instances.
On Windows, the new UUID is extremely likely to be unique on the same or any 
other system, networked or not."

So, that should ne get used on Linux I think.

Maybe it would even more save (though when the Wikipedia Info is correct that 
is not needed)  to just use such a UUID as machine ID  and then create URIs 
like nepomuk:/[EMAIL PROTECTED]/   . 

DanielW


_______________________________________________
nepomuk-kde mailing list
[email protected]
http://lists.semanticdesktop.org/mailman/listinfo/nepomuk-kde

Reply via email to