Charles asked "A buddy of mine asked me if I knew anything about Model
204. I've never run across it, but it sure smells like a cousin of the
multi-value, or perhaps XML, model. Anyone know anything?"
I studied Model 204 in 1975 in my "Comparative Database" class. It was a
new approach to handling data for it could handle both sequential and
random reads/writes. Having mastered INDEXED SEQUENTIAL, it allowed for
immediate placement into a sequential file, whereby you could continue to
read and process sequentially. This was a boon for accounting and
scientific applications alike for you didn't have to read thru the whole
file. (Whooo Whooo Whooo).
Eventually, I programmed in it for about a year in my work study program
for NOAA (National Oceanic and Atmospheric Administration). The project
was super cool for they were flying planes thru heavily air polluted areas
with lasers mounted to the ends of the wings. The lasers had light
sensors that picked up the reflection of the laser light off of the
particulates in the air. We took the information from magnetic tape and
loaded the values into massive files used to track and estimate the
dissipation of the particulates.
I never considered Model 204 as a multi-value variant due to it's rigid
(inflexible) structure. MV's strength lies in the ability of variable
length fields and to change any field to a MV field. As a strong
competitor to IMS and CICS, it did gain some hefty clients during it's
time, however was completely eclipsed by the high availability strengths.
An example of the strength of CICS is it's ability to change stuff within
the core run engine without taking the DB down. Yep, you discover a bug
in the query function in the run engine, and you can replace the code
section for the query function on the fly to fix it!!! This lead to many
shops moving away from Model 204 to IMS or CICS.
A long time ago, there were people who compared COBOL & Model 204 data
structures to those of the MV world. When defining the data structure in
COBOL & Model 204, there is an OCCURs clause that allows a field to be
repeated (at a fixed length) for "X" number of occurrences. While this is
really stretching the definition, it did allow for fields to be repeated
up to a predefined number of values.
I almost laugh ever time I read about the XML capability of any database.
All it means is that it "Can" map to a logical / physical data store. But
believe me, if you have to restrict the length of a field, and it can have
only so many values before it blows up, is not really XML at all.
Please remember that when IBM embraced XML, U2 had the fastest, full
function system in all of IBM.
UniData then developed COBOL Direct Connect (CDC) to solve the reporting /
data store problem. CDC is still available within UniData. We need to
sell that more.
Steve
Stephen M. O'Neal
U2 Lab Services Sales Specialist
Information Management, IBM Software Group
---
u2-users mailing list
u2-users@listserver.u2ug.org
To unsubscribe please visit http://listserver.u2ug.org/