With the fast-paced world of
computers and AI, it is hard to
believe that my first encounter with
computer was in the form of a small
minicomputer – only a server unit with
paper tape as its input! This was in early
1980s while I was pursuing my engineering.
It used to take assembly instructions on
a paper tape as input and give its output
on the paper tape itself. No display, only
few red LEDs blinking very fast to show
that some vigorous computing was being
done! It was obvious that we lost interest
soon enough after trying out a few add/
subtract types of programs. However,
the final year of engineering turned
out to be more interesting since we got
to experiment with microprocessors;
they were the real introduction to minicomputers
and laid the base to try out
many things. As a final year project, we
developed EPROM programmer using a
Zylog microprocessor kit.
In 1982, I joined IIT Bombay for MTech.
There I got a lot more exposure to
computing as well as large computers. The
computer science department had the EC-
1030 Russian computer – 3rd generation
LSI technology (equivalent to IBM 360).
I still remember that this huge computer
used to be in a caged room where only
a few privileged people (admins) were
allowed to enter. We could hardly see
this computer from outside – we were
allowed only till the reception room
where we used to go submit our card
decks as programs (in Fortran). We used
to come to know whether the program
was successfully compiled and executed
only a few days later, as there used to be
a long queue. So, it was very important
to review the code (punched instructions
on the card) carefully so that we would
not have to visit many times to get it
executed successfully! In fact, we used
to feel embarrassed to even visit twice
We indeed have come a long way...
Uday Ramteerthkar
to execute a program successfully. So
were the days, life was tough. It is really
astounding that computers have come
a long way and gotten converted into
laptops, Ipads and realise how much the
current generation is lucky to view the
programs and outputs.. on a nice GUI..!
After completing MTech, I had a choice
to join either software or hardware
company. My role model advised me to
choose software as he envisioned more
opportunities and a successful future
in it. Obviously, I joined a software
company, with the hope that I would
learn both hardware (on my own!) and
software. Initial days were hard, not
very encouraging because I had to work
with the same type of computers with
card decks interface. But, once we got
allocated into system support group,
we had an opportunity to work with
sophisticated systems and software. We
were using Burroughs computers with
DMSII network database and ALGOL
as programming language. We were
supposed to maintain and enhance
ALGOL based generators for report
and edit programs. It was an extremely
challenging assignment as the generator
software in ALGOL and macros ran
into 4-5 thousand lines and was tough
to understand. But the concept of a
specification driven generation was
intriguing and that remained my source
of inspiration throughout the rest of my
software career of 35+ years!
In 1986-87, I got an opportunity to work
on the state-of-the-art platform and
software. Our team developed a CASE
(Computer Aided Software Engineering)
tool on IBM 3090 with DB2 – one of the
first few commercial relational databases
then. In fact, this tool was already
developed on Burroughs platform with
DMSII network database. Our job was
to migrate this to IBM platform and
DB2 with many more enhancements. I
would say this software was well ahead
of its time. It had support for many
modelling techniques (Data Modelling,
Data Flow diagramming, Extensible
modelling framework etc), coupled
with reverse engineering modules to
create models from legacy code. Later
on, this became a foundation for many
tools for model driven development,
maintenance as well as reverse
engineering platforms. In fact, we could
see its major utility in resolving Y2K
problems during the century turnaround.
In 1990s, personal computers and
desktops (on intel 386, 486, Pentiums)
came in a big way that shifted power
to individual programmers. Object
oriented programming, Graphic User
Interfaces became a way of developing
applications with more complexity at
the same time with smooth and easyto-
use user interfaces. I remember that
the Super minis/micros (AS400, DEC..)
were getting replaced by intel-based PC/
Servers with more computing power.
STAY UPBEAT
126 Pensive E-Magazine 127
Migrating to Windows, Unix, GUI
based applications with C, C++, Java
as application programming languages
became the main theme for developers
around the world. While we continued
to enhance meta model-based tools,
program generation for Java, GUI,
object-relational database access layers,
we could see many 3GL, 4GL tools
were mushrooming in the market to
improve the developer’s productivity.
Next 10 years saw a focus on developing
large and medium size applications
to address most of the domains.
Subsequently, the development pace
decreased while the focus shifted on
maintaining and enhancing existing
applications in a robust and consistent
way. Huge data centres with all kinds
of applications were set up for large
organizations. The software services
companies started focusing on providing
cost effective solutions to manage these
centres efficiently and cost effectively.
Meanwhile, parallelly during 1990s,
Artificial Intelligence (AI) also emerged
as one stream to handle complex
reasoning problems. Prolog, Lisp based
modules were developed to handle rulebased
functionality. Somehow this theme
did not catch up fast and remained
dormant for some time. But now that
the compute capacity is increasing
multi-fold… handling of large data (Big
Data), associated analytics and reasoning.
using neural networks has become
very much possible. AI has picked up
significantly and machine learning /
deep learning have become state-of-theart
technology now, to resolve problems
& recommend solutions in many
domains – such as Health / Medicine,
Space research, Utilities, Infrastructure,
Finance etc. For tackling large data as
well as complex rules, parallel processing,
GPUs are also being tried by multiple
organizations
I had an opportunity to work on a
framework in recent times (2015-20)
which combines model-based generation,
analytics, AI techniques to automate
many of the mundane tasks as well as
forecast/predict the behaviour issues and
recommend appropriate solutions..
Currently, the cloud computing era is in
vogue. Organizations are moving towards
clouds to have agile, cost effective
operations as well as to avail latest tools,
software, large data processing capacity
etc. Future for computing sees immense
possibilities. AI, Data Science, and cloud
computing will definitely be in the main
stay at least in this decade.
As I look back, I consider myself very lucky to traverse
and experience through all types of computing systems,
databases, programming techniques and new paradigms such
as model driven systems and AI. It has been a very intriguing,
interesting, and enlightening journey all along!
STAY UPBEAT
Comments