Re: Observability (and yanking thereof)

[Jon Paris <Jon.Paris@xxxxxxxxxxxxxx>]

On 11-Jan-08, at 11:29 AM, midrange-l-request@xxxxxxxxxxxx wrote:

> Yet during the CISC to RISC conversion a few companies made quite a  
> bit of money decompiling code with
> observability removed.

That could be done with OPM code in part because the generated code  
patterns were fairly consistent, and almost everything was embedded  
in the generated code. For example the entire logic for each I/O  
operation is generated which is very different from the "ILE way".

The same folks who wrote those tools expended a lot of time and  
energy trying to do the same thing with ILE code and to the best of  
my knowledge failed fairly miserably.

In part this is probably because the ILE compilers are a rare case of  
IBM eating its own cooking and the generated code makes heavy use of  
procedures in service programs to do its work.  There is also the  
fact that the optimizer (even when no optimization is requested)  
makes changes to the code and, as Mark noted, New MI (or MI prime) is  
not documented outside of IBM.

Jon Paris

www.Partner400.com
www.SystemiDeveloper.com