Re: [OT] System Conversion - An Overview



"HeyBub" <heybubNOSPAM@xxxxxxxxx> wrote in message
news:13ifknbime540d2@xxxxxxxxxxxxxxxxxxxxx
Charles Hottel wrote:

I have learned that we now have around 860 contractors working on
our "new" system. I put new in quotes because the current system was
used for the specifications for the new system. That 860 number is
after they got rid of a lot of SAP people. They tried to replace one
subsystem with 80% vanilla SAP and 20% ABAP, but the vanilla SAP
could not do even 60%. All of this just to say they no longer have
COBOL and now have the new and improved and 20% to 30% slower state
of the art Java. In one lunch room they have a huge chart of system
development processes and paperwork deliverables. If you live in the
USA then sorry, it it your tax dollars at work :-(

This echos what I've learned from long experience.

A large system designed from scratch will not work and cannot be made to
work. You have to start over with a working, smaller system. However, a
large system, produced by expanding the dimensions of a smaller system,
does not behave like the smaller system.

In a large system, malfunction, or even total non-function, may not be
detectable for long periods, if ever. The total behavior of large systems
cannot be predicted.

Some complex systems actually work. If a system is working, leave it
alone.

In setting up a new system, tread softly. You may be disturbing another
system that is actually working.

Somewhat surprisingly, I actually agree with this. While I enjoy change and
large challenges, I think your analysis is pretty much on the button, Jerry.

One of the reasons I get excited about Web Services and OO components is
because you can implement them without the risk of destroying something else
which, as you say, could be working fine.

My experience bears out what you say about starting with a smaller system
and expanding it incrementally. Obviously, encapsulated building blocks
facilitate this.

Pete.
--
"I used to write COBOL...now I can do anything."



.

Quantcast