Lou,

LR> Hi, Kelly.  Your argument makes sense.  I just not 
LR> understand what you meant by "low memory" - which I now 
LR> understand you to mean that swapping is required since the 
LR> apps can not be fully loaded into memory as a exe. 

Right. It's not always the case, but it is possible.

LR> However, since no programmer can know, in general, what the memory
LR>  configuration of a user might be, then that programmer must  presume
LR> that some user might not be able to fully load the  entire exe,
LR> right?  Hence, it seems, based on what you are  saying, having any
LR> .exe write to itself is not a good idea -
LR>  bad procedure, right? 

That was my point, yes.

LR> Hmmm.  I have never thought about the issue of writing self-
LR> modifying code - in memory on a paging system before.  

I don't think it's done very often.

LR> Of course this is very different than having a exe program 
LR> write to it's own .exe file on a HD but it is an 
LR> interesting issue - to me.  Do you have any insights you 
LR> would care to share on this self-modifying code issue? 

Yes... Use a configuration file, or have a separate program write to
the  program's .EXE file while it's not running. Evidently,
self-modifying code  is not possible on OS/2.

LR> I wonder if there is a way to get around this apparent problem of
LR> having a  .exe write to it's own .exe file on a HD?  Based on what 
LR> you said, you have not experimented with this type of  problem,
LR> right? 

Right.

LR> As I understand it, OS/2 is fully multi-tasking, hence it 
LR> is possible for one task to intiate another task to run 
LR> concurrently with other tasks and that sub-tasks can 
LR> intiate other sub-tasks to some level of nesting in this 
LR> manner, right? 

That's right, Lou. 

LR> I do not fully understand how swapping takes place under 
LR> OS/2 - to what extent facilities are available to a 
LR> programmer to control this swapping process as say there is 
LR> under DOS via overlays where the programmer has control to 
LR> determine what is in memory or not. 

Apparently, there is a way for a program to control that function, as
evidenced by the  C/Set++ compiler. No matter how much real RAM you
have, the swapfile grows, as if the program is intentionally putting
stuff in the swapper.dat file, instead of using  RAM. I don't know how a
program would do this, and I don't think that most programs need to
worry about it. It's an OS function.

LR> Since some programmers apparently feel there is value in 
LR> having their exe modify the exe copy on HD I wonder if IBM 
LR> has specifically addressed this issue and may have come up 
LR> with some answers or alternatives. 

.INI files, perhaps?

Kelly

___
 X KWQ/2 1.2g NR X This tagline stolen by KWQ Mail/2.

--- Maximus/2 2.01wb