India's Software growth/debacle

[vamsi@siliconcorp.com]

---------------------------------------------------------------------
Please help make the Manifesto better, or accept it, and propagate it!
---------------------------------------------------------------------

"Dr. Subroto Roy" wrote:

> Is "the client's existing applications" something being recorded in the
> accounts of the Indian software company as an input or a liability or an
> expenditure?  So for example, the existing application is recorded in the
> accounts as worth $x million, to which $y million value addition occurs by
> way of e.g. new code being written in India?

A brilliant point by Dr. Roy although there is a much simpler explanation or a
way to measure software development and I am almost certain that is what is
used by the US and Indian Government Agencies or their contractors.  I don't
think the problem with these predictions is with accounting, however.  If it
was an accounting error then the current number would be more than $5 billion
for this year alone which is not the case.

I will take Intel and Microsoft to make my case.  For example, regardless of
how much development is done at Microsoft its sales are very closely
constrained to the sales of Intel Semiconductor products (and possibly a few
other processors).  But Intel's production itself is bound by natural limits
of semiconductor manufacturing - so many chips per year is the maximum.
Hence, an easy way to predict Microsoft's growth is to take Intel's capacity
addition (and other processors) and apply correlation factors to each that is
consistent with the recent history.  Note: since there are sufficient number
of players in the motherboard/hardware market they are not as relevant.  In
fact, this is exactly how it is done!

Now, for us to predict Indian exports as a whole we have to throw a few other
things into the equation - computer sales, number of IT institutions
(graduations), Internet Bandwidth, power sector, etc.  What is of concern here
are the hare-brained numbers that Government of India comes up with when
estimating simple things like Internet Bandwidth and computer sales.  For
example, the Internet Bandwidth is predicted to be around 400 Tbps by the end
of this decade (or was that 40 Tbps?).  Anyone having the slightest idea about
what these numbers mean would be sick to their stomach!!  For instance,
Internet Bandwidth can be predicted simply by taking the number of ATM
switches installed and applying a correlation factor to it.  Again, there is a
physical limitation to the production of number of ATM switches as is the case
for Intel Semiconductors.

My take on this is that the Indian Government has done a bad job in
calculating primary factors just as Internet Bandwidth, computer sales, etc.
Hence, any other predictions derived from such primary factors cannot be
trusted.


Sincerely,
Vamsi M.


--------------------------------------------------------------------------
This is the National Debate on System Reform.       debate@indiapolicy.org
Rules, Procedures, Archives:            http://www.indiapolicy.org/debate/
-------------------------------------------------------------------------