Session 2601 - ISPF Today and Tomorrow

February 22-27, 1998

This was a presentation on the latest ISPF efforts. Dan Blum is ISPF's new product manager, and has had the job for about a year (he took over from Gina Poole as best as I can remember).

The big addition to ISPF is the "Application Server", and a corresponding client applet written in Java. You use a plain browser to go to a web page that contains the client applet. The applet connects to the Application Server, which starts up an ISPF session on the OS/390 system. After that panels are displayed under control of the browser. Anybody with access to the Internet can be given access to one of your ISPF applications.

This all comes at some cost however. The applet is about 70KB of code, which might take awhile to download. It also requires that each user have a TSO address space created, and that you be running the Domino GO webserver. It can't be very speedy.

The other big enhancement is the new Visual Age for ISPF program. Using VAISPF you can use drag-n-drop techniques to create ISPF panels. The ISPF developers think that this is a big improvement over your having to learn the ISPF panel language, but I have my doubts. First, you still have to code the )PROC section - no automatic generation is possible. And you have to find some way to upload the generated panels to the host, being careful not to introduce ASCII-EBCDIC errors on the way.

There have been some cosmetic changes and minor functional enhancements to the panels themselves. As an example, a developer said they had consolidated the 3.2.A and 3.2.M panels (which are used to allocate unmanaged and managed datasets, respectively). I guess this was a good thing, but it's hard to get excited about it.

Blum assured the audience that his staff was pleased to support their users and that if we so desired that fixes to reported problems would be produced cheerfully and quickly. But, he added, this takes resources that he could spend elsewhere. He suggested that we consider allowing level-2 personnel to defer producing a fix for an APAR for the next release (known as a "FIN", or "Fixed If Next" resolution). This apparently cuts down on some of the bureaucracy inside IBM.

After the session I pointed out that Fixed-If-Next works fine until IBM reversions the product, with a corresponding price hike. Blum shook his head up and down, and told me that he had heard that from many customers and that he was sensitive to the issue.

Back to session index
Back to index of SHARE meetings
Read the disclaimer