SV: Development tools (was Is PowerHouse DEAD?)

[Jon Kvisli]

Thanks for the many interesting and new perspectives of this weeks
debate. At the end of the week I have noticed two of them that I wish to
comment on. To make it look short, readable(?) and easy to skip, I split
it I two postings. :-)  Here is the first: 

Robert Edis told this story of a cost/benefit decision: 

"The CIO of the company wanted everything to be HP/Oracle so the CB
analysis was ignored (shredded, archived, whatever).  None of the cost
savings the CIO touted as a reason for the conversion eventuated so the
company spent millions just to stand still!"

At first glance this may seem to be a silly thing to do. However we are
working with technology that develops and improves over time. While
technology is rushing forwards, anyone who "stands still" is actually
dragging behind. (Pilots will probably call this "negative
ground-speed".) The costs described by Robert is probably spent based on
a wish to keep up with technology. Functionally, it way seem like
"standing still", but the costs of not doing such a conversion may very
well be much larger if the old technology is no longer supported or if
it technically limits further development and the use of new tools. Such
costs are often not considered in cost/benefit analyses like the one
mentioned.

In 2000-2001 my employer spent app. 17.000 man-hours + 6 mill. NOK for
hardware and Oracle licences, on converting two large PowerHouse systems
from HP3000/MPEix/TurboImage to HP9000/HP-UX/Oracle. Very little new
functionality was put into the systems during this conversion. To
end-users and top-level management, this process probably seemed like
"standing still". However, if it had NOT been done, we would have been
stuck with hardware and DBMS that limited further development strongly,
and that is going to be desupported by HP in near future. The conversion
also opened the technical possibilities for using new Windows- , Java-
and Web-based tools to develop new functionality based on the old
systems and databases. We are now doing this gradually while still
maintaining (but reducing) old PH code. 

Critical readers will probably point out that these costs could have
been spent on developing or buying completely new systems. The reason
for not buying is given in my next mail, and the systems was considered
too large, complex and critical for a complete "big-bang" rewrite.

Regards
Jon Kvisli
----------------------------------------------
Principal Consultant
Lindorff Applications as
Hellandtunet research- og businesscentre
P.O.Box 4, 3833 Bo in Telemark
phone: +47 35 06 15 71
fax: +47 35 06 15 01
mail: jon.kvisli@lindorffapplications.com
www.lindorffapplications.com
----------------------------------------------