Razzak,
<10. POOR Database Design and Coding Methods >
I've found this to be true because of my poor design
methods using cursor controlled while/endwhile loops and eep errors in
forms.
I'm
the worst at writing it one way then ending up doing it
another!
Because I found a better way after thinking about what
I'm trying to do and use the code provided to get the best
return.
I've tested many methods of defining var, selecting
data, inserting and updating data.
I had
problems with memory consumption that was due to my coding
methods.
I've
found that doing a lot of var redefinition with in the loop and updating
will consume a lot of memory in NT4.0SP6 Workstation or
Server.
I
changed my methods of what I needed to do and coded
according.
I
use lots of projected temporary tables, insert selects and
do very little var redefinition with in the loops!
This
eliminated the memory consumption (to me not a leak) in my NT
applications!
I've
found many differences between Win95/98, WinME, WinNT and Win2000 (who
knows about XP) on the methods that can be used in combination of coding
versus memory consumption.
<11. NOT taking advantage of
CURRENT versions >
I only
support and write application with the current versions of Rbase, because I've
found that headache after headache will occur on the older version versus the
newer OS's.
In
regard to the Multi User speed test.
I've
found that most of the problems I have are network related.
How
you have the network setup!
You
will have a drop in access time from single user to multi user on the
network!
This
is due to the network setup and the way you configure
Rbase!
You
can not confuse the single use mode versus the multi user mode, the speeds can
not be the same.
I've
tuned a application on a NT network and the speed on accessing data from all of
the workstation is the same across the board in multi user
mode.
But on
the server running Rbase it is much faster access because it's on the local
drive.
If I
access the database from the server to one of the workstation (copy the DB for
testing to a workstation) the access time is the same as the workstation (in
single user or multi user mode)! Different commands will be faster
than others again a method of coding can cause slow response time! I've
done that by the way I select data for lookups and updates versus insert
select. I can do it two ways and get difference time durations for each
method!
KISS!
(keep it simple and stupid)
I've
worked with RBTI on many issues and each time I have come away with a better
understanding of RBASE and it proper use.
I've
had no formal training but I've hacked my way to a more efficient use of Rbase
as I get a better understanding of the different commands and what combinations
give you the best results.
I have
found no other database application that I can do the same with and have
the reliability and repeatability of data and application processing as with
Rbase.
Just
my two cents!
Best
regards,
Oma
