Per (e-mail address removed):
They dont even teach MDB in India.. is it not obsolete?
To the extent that the really large projects get outsourced to India and small,
department-level apps tend to be developed in-house, it might be obsolete in
India.
..MDB back ends are definitely not industrial-strength/massively-scalable
solutions and .MDB front ends don't have the cache' that .NET front ends do.
But both are significantly cheaper to develop than .NET front ends/SQL Server
back ends and, in spite of what some IT people say/think, they do an acceptable
job at a certain scale for much less money. *MUCH* less money....
To wit, I have developed two applications that, some years later, were taken
over by IT and "done the right way".
The first one was a charitable endowment manager. Somebody with a lot of money
wants a tax break right now, but doesn't want to allocate the money until later
- so they set up a charitable endowment fund that acts like a tax exempt
charity. They get the tax break as soon as they turn the money over to the
charity and they get to decide how it is given out later. I wrote a system to
keep track of donors, donations, and allocations and to serve as a tool for
recruiting new donors.
I delivered that application for a tad over $71,000. The user was happy with
it for several years. It did everything they wanted and they got enhancements
quickly and painlessly.
Then one year a server got goofey and the app started going down in the middle
of the day. Moving to a different server would have fixed the problem in an
hour or so - but the company had everything so locked down that it wasn't an
option until almost a year of constant problems had passed.
So, the action taken was for IT to take the app, use it as a prototype, an
re-develop it the "right way" using and Oracle back end and a .Net front end.
Total cost for the clone/rewrite: over two million bucks. No additional
functionality.
The second one was a bond trading application that I developed over a period of
about seven years working with a mutual fund's bond trading desk. Total billed
hours were a little over $225,000 and, over a period of nine years the only time
it went down was when somebody pulled a cable in the LAN closet - and it was
back up and running within the hour.
Upper management decided that several other trading desks should also be on such
a system - but they elected not to enhance the MS Access app that I wrote.
Instead they had IT rewrite it from scratch. I participated in that process
for about two years. There was additional functionality in the replacement,
but no more than 40%. Last time I looked in on it after about five years of
IT's working on it, it was going to be delivered sometime in the next year and
the sunk cost so far was over 24 million dollars.
The original users expressed consternation that they had to use the new system.
SO: bottom line, MS Access applications may not be industrial strength - but
they certainly cost the user significantly less and, at a certain level, can do
everything the user wants.
You wouldn't buy a three-axle dump truck when a VW Golf would do the job at hand
faster and cheaper.