I think the basic concept here has very little to do with any kind of
low-level tweaking or variations on versions.  Let me explain by example:

Software I used to work in had a financial ledger for material
transactions that could get to be VERY large.  If I wanted to run a
Account analysis for a particular period it was always faster to do an
initial select based on date (particularly if you indexed that field :) ),
then doing the additional selects and sorts on the lesser record pool.

My rules were to try and do non-sorted selects first on fields that are
indexed and/or do not involve wild cards.  Then do the any additional
selects and the sorting in a stacked statement.

Rich Taylor | Senior Programmer/Analyst| VERTIS
250 W. Pratt Street | Baltimore, MD 21201
P 410.361.8688 | F 410.528.0319 
[EMAIL PROTECTED] | http://www.vertisinc.com

Vertis is the premier provider of targeted advertising, media, and
marketing services that drive consumers to marketers more effectively.

"The more they complicate the plumbing
  the easier it is to stop up the drain"

- Montgomery Scott NCC-1701


-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED]
Sent: Thursday, November 11, 2004 2:01 PM
To: [EMAIL PROTECTED]
Subject: Re: [U2] PICK Assembler Language

With one exception, I agree with Charlie. Just because the hardware 
speeds allow us to write crap doesn't mean we should. There is, of 
course, a limit to cycle-tweaking. I am not going to spend my time 
looking for the fastest way to prepend a floating dollar sign, for 
example. If I have a case of a large number of transactions triggering 
secondary reads (code file lookup, for example), I will certain take a 
few minutes to 'ask the system' if a dynamic array cache with locate 
would be worthwhile.

The exception I mention is "done once in the development cycle." What 
was efficient at release x may not be efficient at release y. Thus I am 
a big proponent of constant refactoring.

-- 

Regards,

Clif

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
W. Clifton Oliver, CCP
CLIFTON OLIVER & ASSOCIATES
Tel: +1 619 460 5678    Web: www.oliver.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~



[EMAIL PROTECTED] wrote:

>Just because we have a quadrillion microseconds to play with, instead of
a  
>50 millisecond timeslice, doesn't mean we have to waste them. A few
moments to 
>consider performance, done once in the development cycle, will pay
benefits 
>every time a more efficient program is run. I have to believe that  Adm
Grace 
>Hopper is smiling down on you both.
> 
>Regards,
>Charlie Noah
-------
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2ug.org/
-------
u2-users mailing list
[EMAIL PROTECTED]
To unsubscribe please visit http://listserver.u2ug.org/

Reply via email to