On Sun, 21 May 2006 18:04:32 -0500, Miguel Bazdresch wrote:
database includes an entry for every uid. That's quite an inefficient
design, if you ask me.
There are arguments for and against...
If you have high uids and no sparse file support it's a disaster!
But on the other hand it can be argued to be very efficient for access
because the
records are fixed length, so you just muliply record length by uid and lseek
directly to the
record you want. It's like the utimate hash-lookup with no collisions! Your
access is O(1). And there's no possibility of getting duplicate IDs in the
file nor overhead to maintain uniqueness of the ids.
--
http://linuxfromscratch.org/mailman/listinfo/lfs-chat
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
