Thursday, July 19, 2007

Mainframe Journal

I'm starting to round up for my holidays, so I guess this is the right time for some reminiscing. In my former posts on the 'Introduction to the New Mainframe' series, I pointed out how difficult it was in the late 80ies (when I set my first steps in the mainframe world) to find decent ànd comprehensible information. I see that the Mainframe Weekly blog is also writing a similar post on that item this week.

One source I forgot to mention and one I'm still very grateful to was the Mainframe Journal - which later on become Enterprise Systems Journal (ESJ). So when I received a mail of Bob Thomas, commenting on my blog (I hope you don't mind me mentioning this, Bob), I actually went to my attic to dig up my old issues. I still have a clear picture in my mind of some of the (really) artistic covers, like e.g. the Dali like cover with the melting watch of the August 1990 issue.

Paging through some issues I also stumbled across the June 1994 issue with the tempting cover : 'Nine Predictions'. Well we're 13 years later, what happened to those predictions written by Howard Fosdick, who wanted to predict the situation at the end of the (previous) decade ? I'll just let you judge for yourselves. One of them was :
Visual Programming on the Desktop.
The current trend toward visual program development of PC applications will continue. the percentage of programs that do not reside on PCs or are developed outside of visual programming will decline. Few companies will convert existing applications or junk their mainframes. The ongoing desktop development of new applications will gradually move I/S into the world of the PC interface over the next several years.
Quite true, no ?
Another one ?
A Time To Squander Hardware
It may be a trivial prediction (but I want to get one right!) - as companies evolve towards C/S networks, the point analysts have predicted for years will finally be reached - I/S will squander hardware to cover deficiencies in performance analysis and optimization. Rather than nderstanding how to write effecient programs and build high-performance systems, many companies will just add servers. Networks of underutilized and poorly optimized servers will exist in large companies as overworked I/S staff hustle just to keep them alive. After all, they come cheap and interoperable (the servers, that is).
A paradoxical but simultaneous trend will be the maturation of software for the administration, performance monitoring and maintenance of C/S networks. But the expertise to comprehensively optimize systems with these tools will be missing at many companies. Most will prefer to pay less for expertise, hiring low-end employees, and pay more for hardware.
Visionary, no ?

It makes me want to repeat the video I posted some time ago and that popped up at the IBM-Main list a couple of days ago :

No comments: