Re: [bug-recutils] GSoC: Ideas for Recutils

[Dmitrii Kashin]

Cartainly, it's a very good idea to create separate index file.
I thought about the same way of solving.

But I don't think hashing (checksumming) will be slower than query by
plain index column. This depends on database size. If it is very big, so
you can use hashes to bind rows with different index files. Algorithm
complexity for index search is proportional to log(N), but it's not a
good idea to keep 100 MiB of indexes in memory just to find necessary
one. This practice is used by some filesystems I heared.

Well, I see you want this project more than me. Get it, guy. There's no
competition here. This will be good practice for you. =)

PS: It will be a good idea to have upstream maintainers to use git or
some other cvs you want. IMHO.

-- 
**************************************
*  jabber:  address@hidden   *
*   Registered linux user #546240    *
**************************************