RE: Call back procedures

Performance:

Are you thinking of the system performance? Are you asking if a callP to procedure is different than a callP to a procedure pointer? Like Scott said, at compile time you won't get the parameter validation, but as far as the actual call goes, I'm guessing that under the covers it's always doing a procedure pointer call (someone call me on that if I'm off base).

Or are you thinking of the logic performance? Like I suggested the potential to call the error procedure multiple times (multiple warnings per record (I presume)) with the thought that you can leave processing if an error occurs vs one call at the end of processing? This is where I was saying do what's right. I mean if your team has agreed to accepting the level of complexity call back procedures introduce (although once you understand it, the idea of it being complex fades away) then I'm guessing they've signed on to the basic idea as to why you need it, right?

Maybe some further explanation on this concern would be helpful.

-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Rick.Chevalier@xxxxxxxxxxxxxxx
Sent: Friday, May 08, 2009 1:48 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: RE: Call back procedures

I didn't mean to overstate the performance debate. There are two concerns, complexity and performance. We are ok with the complexity, I just couldn't come up with a good argument either way on the performance not having any experience with it yet.

-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Kurt Anderson
Sent: Friday, May 08, 2009 1:39 PM
To: 'RPG programming on the IBM i / System i'
Subject: RE: Call back procedures

It seems to me that your idea for using the call back procedure is beneficial. It's smart and thought out. Where's the performance hit? Because you may have 10 warnings and the new way would call the error routine 10 times and the old way would only call the error routine once? What if you have 1 error and 100 warnings, and the error occurred first, boom you're out of your edit without encountering the 100 warnings.

I don't know how can quantify that, one way or the other. Even so, like others have said, it's just like any other procedure call. Personally, I'd side with reason vs performance on this one... not to say there really is any performance hit. Are the people concerned with performance take that into account when they program a subR or procedure call?

On a sidenote, I'm glad to see someone's using (or wants to use) call back procedures. I have an idea for one to put into my shop... but one change at a time. :)


Privileged and Confidential. This e-mail, and any attachments there to, is intended only for use by the addressee(s) named herein and may contain privileged or confidential information. If you have received this e-mail in error, please notify me immediately by a return e-mail and delete this e-mail. You are hereby notified that any dissemination, distribution or copying of this e-mail and/or any attachments thereto, is strictly prohibited.