[346] in Coldmud discussion meeting

re: floating point

daemon@ATHENA.MIT.EDU (Fri Jun 3 23:17:59 1994 )

Date: Fri, 3 Jun 1994 14:03:55 -0700
From: rayn@q.crossaccess.com (Ray Nawara jr.)
To: coldstuff@MIT.EDU
In-Reply-To: <199406031941.MAA04488@netcom.com> (deforest@netcom.com)


> While server-based floating point support would be far faster, it
> would not be exceedingly difficult to code floating point in-db.

Thats what I was thinking, too, but it seems to me that any in-db
floating point package will have quite a bit of overhead that can be
easily eliminated if a new type were implemented. I guess the question
is is the overhead enough to warrent the work the server change would
require.

A server implementation would probably have more precision for less
effort though, as well. (effort on precision, not on total
implementation)