On Monday 03 December 2001 07:26 pm, you wrote:
> I could be wrong, but I think this is equivalent or similar to a scheme
> that was proposed a while ago (perhaps by me - I can't recall).
>
> This stuck with the current DBR system, but when today's edition didn't
> exist it would try yesterdays, then 4 days ago, then 8 days ago, then
> 16, exponentially searching backwards.  When a page is found it
> exponentially searches forwards in a similar manner, then backwards,
> forwards, etc etc (sorry not a clear explanation, I will leave it as an
> exercise for the reader to fill in the details).  In common with your
> suggestion, this should find the most recent edition in logarithmic
> time.
>
> Ian.
>

Searching 1, 2, 4, 8, etc. inserts ago wouldn't work as well.  For the Binary 
Tree DBRs, root branches will be requested more often and then become more 
reliable.  If one of the lower leafs becomes inaccesible, the next-highest 
root would be more relaibily accessible.  This is because of the required 
traversal methods of the binary tree.  Also, a power of two lookup might not 
find the most recent one if it misses a day.  Binary trees handle that 
without too much overhead.

The Binary tree automatically handles change in update frequency.  A user 
could insert two versions less than a second apart without problem, or leave 
it for a year without updating.  It can scale down the time interval 
infinitely small too.

Scott Young

_______________________________________________
Devl mailing list
Devl at freenetproject.org
http://lists.freenetproject.org/mailman/listinfo/devl

Reply via email to