Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!apple!sun-barr!ames!xanth!mcnc!unccvax!cs00chs
From: cs00chs@unccvax.UUCP (charles spell)
Newsgroups: comp.databases
Subject: Database performance
Keywords: performance on low-level access
Message-ID: <1548@unccvax.UUCP>
Date: 10 Jul 89 01:59:55 GMT
Organization: Univ. of NC at Charlotte, Charlotte, NC
Lines: 18

I see that there are many public domain and shareware data libraries
(using C as a DML) out now. CData seems to be the most popular of these
type packages. 

Has there been any benchmarks run with these cheaper vs the $500+ DBMS
packages? It seems that since they all use the popular Btree indexing 
alogorithm and there should be very little difference in performance.
I've also seen Btree+. What's the difference between the Btree and 
Btree "plus" (other than a marketing gimmik) ?

It seems that a good C programmer would only essentially need a few fast
library functions (find, add, update, and delete) to build a database system
of any size. Of course, you could create as many utilities as you want,
but it seems they would all be just C programs with calls to the essential
functions.

Could the performance difference between FoxBase and Ashton-Tate packages
be in the low-level access functions? I find it hard to be anything else.