With all the bad mouthing that poor old VTAM has been copping I will jump to its defence. Some of the most interesting things I got to play with were due to it and GCS. First off we modified our OS/PLI 2.3 source code and added some function to GCS such that we could write apps in proper multi-tasking PL/I. We created some VTAM support routines and we were able to create a system that allowed us to switch messages from our VSE systems to our Series/1 (yes we had these babies too). When we phased out the Series/1s, which were responsible for talking an async protocol to the racecourses around the state to combine pools, we used the DATE support of NPSI to bring that function into a virtual machine. I used to talk to VMSHARE using another GATE-based NPSI application. All this was written in PL/I and served us for years and made the company a lot of money.
Now as far as the joys of NCP generation and VTAM topologies, and SNA protocols in general yes I am happy to live without them now. It reminds me of an ancient joke: John Akers answers the phone: Hello Caller: John Akers? JA: Yes. Caller: John Akers of IBM? JA: Yes. Caller: John Akers of IBM, White Plains? JA: Yes! Caller: John Akers of IBM, White Plains, NY, USA? JA: Yes, WTF do you want????!!! Caller: Just wanted to let you know how it feels to set up an SNA session.
