[RC5] OGR size labelling?

Barry K. Nathan barryn at pobox.com
Fri Feb 25 16:40:28 EST 2000


> There has been a lot of complaints/disscussion regarding inaccurate
> percents on OGR stubs.  Isn't there a way that you could just attach some
> data to each stub basically saying how many nodes is has?  RC5 kind of does
> this, but i know that OGR stubs are extremely varied in size.

The problem is that the number of nodes isn't known in advance. I *think*
that during the previous (non-d.net) OGR projects, there were rough
estimates for # of nodes per stub. They were not incredibly accurate, but
they still helped.

One fact that may put things into perspective: the client code can play a
part in the number of nodes for a given stub. 

For example, one of the clients used in the previous OGR searches, GARSP,
used an external data file called choose.dat. A larger choose.dat file
meant that the client would need to examine (or whatever the correct verb
is) fewer nodes, and thus finish the stub more quickly, but at the cost of
extra RAM usage.

I *believe* that, for at least one of the clients used in the previous OGR
searches, the 32-bit and 64-bit versions of the clients also resulted in
different node counts for stubs.

I hope this email is, if nothing else, interesting. :)

-Barry K. Nathan <barryn at pobox.com>

--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list