I have been talking with Aaron and Tim (and the fog has set in and wondered if Pat, or some other Guru of Bitology, would clear up the Muddy Waters Blues (great album by Paul Rodgers) ..

I believe these 3 are available for access to OI data ..

CallFunction

CallSubroutine

SendScript

My question is .. What exactly is the difference in these routines. Except for Function which obviously returns a 'value' (and can return multiple values during process through send_dyn) and SendScript is much like the commandline in the system editor.

I have been able to use CallFunction and SendScript interchangably (except calling params)

So is there a 'this works best for this' and 'this works best for that'? Process speed differences? Memory used? CPU usage? Future usage when we get MultiThreaded engine

No charts and graphs please .. remember who you are talking to

Thanks

dsig@sigafoos.org

DSigs Radio Free Orgeon

If both of these conditions hold:

1. You only need to return data to the client when your OE routine returns.

2. The client (or more accurately the client thread) making the request must wait for the reply and does not need to perform any other tasks while it is waiting.

Then CallSubroutine and CallFunction are the most efficient way to perform such tasks, particularly with regards to the number of network packets used.

If any of these conditions hold:

1. You need to stream back many "rows" of data back to the client. A good example is if you initiate a selection into the database and want to return the matching rows one at a time as you find them.

2. It is a lengthly process and you want to return status information about the request in progress.

3. The client (thread) making the request needs to perform other tasks in the interim between making the request and receiving the reply.

Then you must use the SendScript family of calls.

HTH,

Pat

Mr Pat ..

kudos and thanks ..

dsig@sigafoos.org

DSigs Radio Free Oregon

View this thread on the Works forum...