home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!usc!cs.utexas.edu!rutgers!igor.rutgers.edu!romulus.rutgers.edu!morrow
- From: morrow@romulus.rutgers.edu (John Morrow)
- Newsgroups: alt.cobol
- Subject: Re: The future of COBOL?!?!
- Message-ID: <Jul.30.21.06.59.1992.1772@romulus.rutgers.edu>
- Date: 31 Jul 92 01:06:59 GMT
- References: <1992Jul30.022405.8953@uwm.edu>
- Organization: Rutgers Univ., New Brunswick, N.J.
- Lines: 89
-
- rca@csd4.csd.uwm.edu (Robert C Allender) writes:
-
- >After reading one of the responses to "The Inevitable Question" I was
- >compelled to ask this of the COBOL programmers in the world:
-
- > The C language has caused a bigger uproar in the past few years than
- > Pascal did way back when. Now there's the object-oriented C++ that
- > uses classes and promotes encapsulation and polymorphism. Since a big
- > problem with COBOL maintenance seems to be the endless patching of
- > endless patches in old code, What do you think about these proposed
- > solutions that C++ supposedly offers?
-
- I, personally, do not think C or C++ will ever replace COBOL in one of
- the primary areas in which COBOL is used -- primarily business programming.
- If COBOL DOES wind up getting replaced in this area, it will probably be
- by a forth generation language or something like Natural that "looks like"
- COBOL.
-
- There are several very important reasons for this.
-
- 1) The sheer volume of existing COBOL programs linked with the current
- "lean and mean" approach that businesses have been taking make it
- unlikely that many businesses will be able to afford a major redesign
- of an existing system in the near future.
-
- 2) Many existing systems work very well. There is even less reason to
- replace a system that works. Remember, the people that often make this
- sort of decision are management, not programmers or analysts. When
- programmers and analysts DO have input, see (5) below.
-
- 3) Many of the applications written in COBOL don't really need the
- features of a "data structures" language like C. Bit level management
- is not necessary for most accounting and record keeping functions.
- Many "data structure" type functions also can be performed outside of
- COBOL programs using either sort utilities or a pre-existing database
- interface such as VSAM, Adabas, or DB2.
-
- 4) Many of the programmers who write in COBOL are not "computer scientists".
- They know no language other than COBOL and would have trouble adapting
- to a "scientific" looking language such as "C". Companies filled with
- programmers hired to program in COBOL would have trouble training large
- numbers of people in an entirely new system. Firing large numbers of
- people and hiring new ones is not only difficult but has a severe
- impact of moralle and causes the loss of large amounts of experience.
-
- 5) Due to (4), many people who were hired to, say, run a mainframe or
- program in COBOL have a vested interest in making sure the company
- CONTINUES to do things as usual. Upgrades are one thing but new systems
- are another. The Wall Street Journal had an article on this a few years
- ago to the effect than many MIS departments resist using micro-computers
- because they are mainframe people and it would put them out of work.
-
- 6) COBOL, in my opinion, is MUCH easier to maintain than C. If I hand
- you a standard C program and a standard COBOL program that does the
- same thing, written by someone you don't know, I bet you could more
- easily figure out what the COBOL program is doing, PERHAPS EVEN IF
- YOU DON'T KNOW COBOL. (I am willing to discuss this point if you want).
-
- 7) The same COBOL code you wrote 20 years ago can always be brought back
- in if you need it. If you toss COBOL and go with C or C++, you could have
- problems.
-
- 8) Many of the existing files/datasets used by companies presently using
- COBOL are formatted with COBOL working storage and file I/O in mind.
- Changing languages could require large amounts of data reorganization.
-
- 9) COBOL is probably more forgiving than C and easier to debug. I assume
- similar statements could be made about C++ but I may be wrong.
-
- In general, the above issues might not be a factor for a small company
- with a few hundred records, a few dozen programs, and three good
- programmers. But when you get into large corperations with hundreds
- of employees, thousands of programs, and millions of records, it
- becomes more of a problem. I have little doubt that COBOL will be
- around for decades -- perhaps as COBOL 99 with new features such as
- recursion and pointers -- but with COBOL syntax none-the-less. Companies
- have invested too much in it to drop it, expecially if their current
- system works which it usually does.
-
- >I'm certain that this will be an interesting debate!
-
- I look forward to seeing what others have to say.
-
- John Morrow
- morrow@remus.rutgers.edu
-
- PS - And remember -- we are mutants here. Many COBOL programmers work
- on IBM MVS/OS machines with no net access. Academia does not
- necessarily represent the real world.
-