Re: local procedures and performance

[Sam Spritzer <SSpritzer@GW.CTG.COM>]

I, for one, do not like to use local procedures, if I can help it.  I have seen some SQRs that are loaded with those and to ask the human brain to do a trace on those is sheer murder.  And to top it all, I didn't see any dramatic performance differences.  My coding standards simply require me to make the SQR as easy as possible to follow by even the most inexperienced programmer even if it 10 more pages of coding.
Just my two cents worth,
Sam

<<< the dragon <ceprn@HOTMAIL.COM>  8/24  6:15p >>>
Hi guys and gals,

Another theoretical discussion..... (but it doesn't involve cats, explorers
or any of that other stuff.)

I am modifying some programs for a client that were left by an sqr
developer.  2 major sqrs and about 15 sqcs.  The programmer of the sqrs went
wild with local procedures.  Personally, I don't use local procedures if I
can avoid it, not because I am not proficient enough to, but because I
beleive in the K.I.S.S philosphy for when I depart and the client has to
support what I have left behind.

In any case.  Local procedures and performance issues don't seem to be
documented all that well anywhere.  Nothing of consequence is written about
them in anything sqribe publishes, and the following sentence is in "Sqr in
Peoplesoft and Other Applications" - 'Others may argue that using local
procedures involves too much overhead' (p345).

OK, while I do *NOT* want to see an argument start about the pros and cons
of local procedures, I would like to hear about experiences where  the use
of local procedures impacts the performance of the sqr - I have one local
procedure that is called and executed about 50,000 times - I want to know if
making it a global procedure will make the program run faster, slower or
have no effect.

thanks,

clark


_______________________________________________________________
Get Free Email and Do More On The Web. Visit http://www.msn.com