Re: Is Ada likely to survive ?

[dewar@merv.cs.nyu.edu (Robert Dewar)]

Marin says

    The stock market is about making money - not making mainframes. So
    I can't really consider the value of IBM stock to be an indication
    of the mainframe market. IBM makes lots of stuff - PCs, servers,
    devices, and, yes, mainframes. I didn't say they were gone - just

There is no secret to where IBM's money comes from -- just take the time
to do some research -- in fact all the material is available on the Web!
Take the time sometime, you will clearly find it illuminating.

    Oh, I don't know about being in a "limited environment" - As I
    said, I work for one of the larger corporations in the United
    States. This trend *could* be isolated to just United
    Technologies, but when I talk to folks at other large
    corporations, I generally discover the same trend - migration off
    of mainframes and more reliance on workstation/servers, networks
    and PCs.

Yes, this migration is going on, but as everyone involved in the IS
industry is very much aware, the idea that mainframes can be easily
replaced by client server models has turned out to be an illusion.

It is possible that UTC is indeed atypical. It is certain that your
ideas do not match the larger picture. So, if you are accurate in
your assessment of UTC (often I find that engineers in companies
have zero idea about what is going on in the IS side of the house),
then I would guess that this is the only possible explanation.

    As I said - I'm not saying "gone" just dramatically declined.

No, noone has seen the expected dramatic decline. You have to look
in the second deriviative to see any decline at all. And even there
the figures are surprising. You are making claims that simply cannot
be substantiated except by wishful thinking.

Incidentally, it is the case that the IBM share of the world wide
mainframe market (which is itself growing) is declining, so for a real
picture of the whole mainframe market you need to look at more than
IBM, and in particular, the Japanese mainframe companies have been
doing very well recently.

    I think the question I was tossing out was "at what point do you
    want to consider a language 'dead'?" Certainly, there are people
    who still use Cobol, just as there are still people who use
    Pascal, PL/1, Algol.

The difference between COBOL (Ada people should try to spell languages
correctly :-) and Algol is that it has a much bigger share of the
market. It is the most widely used language in the DoD today, and
is the target of choice for most 4GL's. Many new applications are
being developed every day in COBOL.

Some engineering and technical people imagine that obviously C++
could be used in place of COBOL, but the fact of the matter is
that the COBOL infrastructure (i.e. COBOL + CICS + all sorts of
other stuff provides an environment for IS development that is
not easily matched by any other language). Ada provides some of
the basic features, but still needs a lot building around it to
make it competitive with COBOL for this application area.

The kind of pronouncements that Marin is making are very common
among technical engineers who know very little about the IS market.
One of the problems in this field is that there is a huge barrier
between the IS and non-IS sides of the house at all levels (most
CS professors know nothing at all about IS development, and it is
really a rather shocking gap, given the importance of this area).
Similarly, I often meet engineers who actually think that all the
world uses Unix or NT, and think that COBOL has died out. Just
yesterday, I was talking to someone in the DoD who thought that
DoD was behind commercial industry because it still used COBOL
extensively, and he assumed that the rest of the non-DoD world
had abandoned it.