On 23 Feb 2004 12:59:37 -0800,
[snip of earlier discussion that's not relevant to this post]
The point isn't that languages go to ZERO use. Undoubtably there are
still poor programmers out there flipping 1 and 0 toggles on a PDP-8.
And people use FORTRAN. And COBOL. and c. Less every year, but of
course the number isn't zero. Never said it was.
No. But you seem to be saying that the number approaches zero, or
becomes negligible, and that's what people are disputing.
An no I'm not kidding that c is virtually dead as a new development
tool. I attend Programming conferences all over the country and I
can't remmeber the last time anyone even mentioned it. What company
would christen a new project in c?? C'mon get your head out.
If you base your opinion on "programmer conferences" then your opinion
is not based on real world experience, and severely skewed. Those
conferences generally center on whatever is "new" or "hot" or "sexy",
or whatever terminology you want to use.
In the real world, there are large amounts of code that took years and
years to develop, and that are still maintained.
Especially Cobol and FORTRAN have a large market-share there. There
are large amounts of _new_ code written in FORTRAN every day. The
language is extremely easy to use and powerful for
computation-intensive tasks, and sports a very large number of
libraries and tools that mathematicians, astronomers, fluid
dynamicists and other scientists can use, and therefore do use. From
my experience, FORTRAN is one of the most widely used languages in
computationally intensive sciences.
C is a language that is still incredibly hard to beat when it comes to
speed, tersity, and small footprint. Especially in the world of
embedded devices that makes it a very well-used language. Major
operating systems in the world are written in C, and are maintained in
C. Many freely available libraries are written in C. A lot of
commercial software is written in C. many real-time applications are
written in C. The comp.lang.c newsgroup is a very busy group. C is not
dead by a long shot.
If you want realtime stats on PROGRAMMING LANGUAGE usage, go to DICE
or a similar site and search on keywords. Perl's stats are actually
disappointing, as are those for java since they are through the roof.
If you want to argue that Perl's usage is on the demise, which seems
to be what started all this, then do so. Don't try to extrapolate from
other languages, especially if you get your facts wrong, and you base
your experience on marketing-driven conferences.
But my point still stands, that developers always need to be looking
at what's next language wise.
I don't agree. Developers should look at which language is most
appropriate to the task at hand, and which development fits in best in
the already existing framework that the code needs to operate in.
You'd be a pretty stupid developer if you decided to develop a new
application for your company in a language when all other code in your
company is written in another language. You'd be pretty stupid if you
tried to write a real-time application, or a computationally intensive
application in Perl. You'd be stupid if you wrote an application that
requires a small memory footprint in Java or Perl. it would probably
not be the right choice to write an application that requires a lot of
text processing in C or Java.
A language is just a language. The compiers or interpreters that come
with those languages give the whole environment properties that you
select for. Always picking the newest and hottest language for
development is as stupid as never looking at new languages that come
along. You simply pick what's most appropriate, and what fits in best
with what's already there.
Maybe in your companies things move
real slowly and you can make a career on FORTRAN, but the hi-tech's
would leave you in the dust. I'm still employed while I can point to
dozens of ex-collegues who are in other vocations now for the simple
reason they tried to make a career out of one or two tricks in their
programming portfolio. The hard truth is that it doesn't work.
This may come as a shock to you, but _most_ companies move really
slowly. If, as a company, you actually do write code yourself, and you
have a serious code base, what exactly would be the business case to
rewrite, test and debug all that code every time a new fad language
comes along? Now and again, an opportunity arises for companies woth a
serious code base to rewrite it all, and sometimes companies do.
However, this invariably requires a large investment, and the returns
on that investment need to be apparent to the beancounters.
Small companies, with little code and small development groups, are
more likely to now and again change the platform they work with,
simply because the dynamics of the IT economics in a company like that
are entirely different.
Apart from that, there is risk management to take into account.
Whenever software failure can cost lives, or whenever it controls
money, the amount of testing and quality control easily costs many
times more than the actual development time. Testing and QA are much
less sensitive to the language used than development time. The
business case for throwing away well-tested code for a new code base
that needs to be tested from scratch is going to be very shaky.
Again, you are basing your opinion on a small subset of the real
world, instead of realising that there may be a large world outside of
your experience that you simply don't know about (and you don't seem
to know about it, judging by your assertions here).
I guess the argument is that there will always be code to maintain.
And that's true. If you want to be a code maintainer have at it.
Someone has to do it.
That is one argument, but the argument is also that there is a lot of
new development in languages that you call obsolete, because other
people use different criteria to choose their language than you do.
The "newness" of a language is not something I've ever heard a good
programmer or IT manager use as a criterion for selecting it. I've
seen bad IT managers and programmers select new languages based solely
on marketing hype, and I've witnessed several projects go down the
gurgler because of that.
Martien
rogramming portfolio. The hard truth is that it doesn't work.