Cobol - Yuk!

ok, so there r a thousand or so jobs on cobol, but i seached the same site for all the main languages i could think of and heres what i found:

c++ 10249
Visual Basic 7290
Java 8704
perl 2149
HTML 4982

cobol 1092


They ALL have more jobs than cobol. And this is not the only reason why its pointless learning cobol, heres a few more for you:

A) It is a HORRIBLE language to program in. Everything has to be lined up perfectly, in the right colums, you aret even allowed to go past colum 72.

B) It takes ages to program in, everything is massive words to do the simplices things. For example, most languages can set a variable to 5 quite easily, just by saying var1 = 5 , but in cobol that would be move 5 to var1.

C) It breaks all the standards from all other languages for no reason other than to be different, for example, arrays are not arrays, they are table, procedures are not procedures, they are sections etc.

D) It uses a lot of GOTO's making it a B**** tring to degug.

E) All cobol compilers i have ever used are very itatating. The one my college used cant even scroll. You press the down arrow and it scrolls about 150 lines/per seconds. They give stupid hard to understand errors, like for instrance, say you miss a full stop and you try to compile, instead of telling you you have missed a full stop it will say something along the lines of "A scope delimiter was not used correctly".


If u like, i can go on to reasons f,g,h all the way to about zzz? But my point is, its a nasty language which is why it is in a rapid decline!

I would love to hear all the reason. Thats the only good thing about cobol, Its fun finding more reasons why it sucks:)

F) There are about 5 different ways to do absolutly everything, so u think you can program in it and then you go to read someone elses code (to copy it) and you dont understand it at all.

G) It behaves irratically, you will get an error in the program and whilst you are tring to track it it will suddenly seece to exist and then just come back the day b4 the deadline, even if you hav't changed the code at all.

H) Every cobol programmer loves dot matrix printers because appartently dot matrix is the only type of printer you should ever used with cobol because it doest work properly with other ones.

I) The editor i am forced to use (and probably every other editor) :

i: Has irrataving old dos shortcut, like shift ins to paste
ii: Automatically saves everything without asking you.

J) You can only run programs if they are in the root of the drive they are saved on.

K) When you compile programs, it creates about 5 other files in your root directory, all with different, but undefined file extenstion, leaving you with a huge mess of files.


Cant be bothered to write anymore, but im sure you get the gist of how crap it really is.

It is a Mainframe language.

Nobody in their right mind would use it on a PC, but for Mainframe use it is not terrible. VB and C are not used in the mainframe world.

PL/1 is a better mainframe language, but not as widely used as COBOL.

If forced back into the mainframe world, I would prefer PL/1, but would tolerate COBOL because there are more jobs available for COBOL programmers.

It might not be bad to learn COBOL since you could end up working fro a company that used both PC's and mainframes. You would be worth more money if you could program for either type of computer.

The worst of the mainframe world is not COBOL, it is the overall architecture of the man-machine interface.

J. Davidson: Most of the fortune 500 companies still use them. For dealing with really big data processing jobs, they still have their uses.

I would hate to do utility billing for 25 million customers on a PC using Visual Basic. Even on a network, some of the mainframe data processing jobs would be formidable.

Mainframes might be getting there, but they are not quite obsolete. Until they are, COBOL, ugly as it is, will be a necessary evil. Again, I prefer PL/1 over COBOL as a mainframe language, but it is not as widely used. There is no doubt that VB is more fun than either.

The days of the mainframe are numbered. Whats the point in learning the skills for a job that i will loose in a while (ok so it might be 10 years but thats nothing in the sceme of things) when i can earn more money doing something more fun that will be around for longer. And no, they would do processing for millions of people in VB like someone back there said, because VB is a RAD for Applications, and its very slow. They might however, use C,C++ or C#.

Well they wouldn't use C#, that's just M$'s version of Java with a new label stuck over the original one.

Anyway why learn COBOL just because it's still got some life in it? Sounds like a very weak argument to be learning it to me. My guess is that it will last in industrial terms just about as long as those who are trained and have experience in it are still around, so rather than making their jobs even scarcer why not go and learn something more useful?

I will correct myself, i do know a company that still uses mainframes and cobol. My sister does cobol programming for them and makes as much as me with twice the experience. She never took a cobol class in school and got hired and trained. They are phasing out the mainframe though to pc based, although very beefed up pc with oracle db.

Mainframes using COBOL will surely last for at least ten years in the US, if not elsewhere. I expect them to last at least 20-30 years due to the horrendous cost of rewriting billions of lines of code, some of which is not well documented. In addition to rewriting code, there are huge costs associated with either firing & replacing existing programmers & staff or retraining them. It may not cost much to fire 10% of your staff when downsizing or economizing (some paperwork & severance pay), but it costs a lot to fire and replace large numbers of employees or to retrain them in new technologies. Furthermore, there is a lot of procedural work external to the computer department which has to change if you switch from a mainframe to a PC network, involving additional costs in retraining people.

The computer industry and those using computers have been plagued with forced inefficiencies for 30-40 years due to problems with compatibility, reprogramming, and retraining costs. To some extent it is like having a tiger by the tail. It may have seemed like a good idea when you grabbed hold, but how do you safely let go? Mainframes might be obsolete now from the point of view of a company just starting up, but they will survive long past their natural life in existing companies due to compatibility and retraining problems.

Some startup companies might erroneously buy a mainframe when a PC network would be better just because that is the way other companies in their industry have been functioning for 50 years. People seldom learn from the current state of the art knowledge available. You folks probably have no concept of the stupid design errors plaguing the PC World, because the early experts paid no attention to the mistakes made in the mainframe world 50 years ago.

For the last 15 years of my career in the mainframe world, I made at least 20% more due to a knowledge of PC hardware and software. In addition to the extra bucks, I got away with being a bit of an individualist (sometimes being insubordinate) because they would have to replace a mainframe programmer, as well as finding a PC hardware tech, software installer, PC user trainer, and a PC programmer. 90% of their work was mainframe based, and they only had 3 full time PC people in a department with about 200 employees. One of the 3 was a manager for the other two. I actually knew more than the two PC workers and took care of the things they did not understand.

For the next 10-30 years, I would expect a modern PC expert with knowledge of COBOL & mainframe work to able to find situations like the one I found, as described in the previous paragraph. Knowing two disciplines is always advantageous when compared to knowing only one.

Hear hear Guv, wise words. The impression I've gained is that a lot of the posters in this forum haven't been to the big end of town and seen enterprise level computing, by that I mean organisations with thousands of employees.

In the large enterprise/government department mainframe rules. I am a windows analyst/programmer but my first job was as a VAX/VMS operator for 2 years and I have a special place in my heart reserved for VAX\VMS running COBOL because it is virtually bullet proof ( and night shifts on a big site are a riot!) I also charge a premium because I know mainframe inside out, when things go wrong, time on mainframes has made a big difference for me a number of times in timely fault resolution.

I think that the people decrying mainframe and COBOL simply haven't been exposed to major enterprise level computing.

Yes, mainframes clearly have a strong future, why didn't I see it before? I expect all the companies that are phasing them out are just giving them a rest :rolleyes:

I have been to 'the big end of town' as you put it, although briefly, and mainframes are certainly not seen as something to be held on to and cherished where I was. The old mainframe had been hauled out a few years prior to my arrival.

Quote:

---

Originally posted by lord_dude
Can anyone that knows what their taking about settle an argument for me please.

I am being forced to learn cobol against my will by my lecurers who say that cobol is the best language in the world. I think it is the worst. If anyone could replyfor me so that i can show my aginglecturers how rubbish it is now, i would appriciate it.

---

It couldn't be that your lecturer could not get a job programming COBOL any more but rather than just giving up on it decided to reuse the skill by teaching it instead ;)

Ksin: The comments I made in a previous post related to reasons for learning it, not reasons for liking it or considering it to be a good language. Compared to PL/1, Algol, VB, and a few others, COBOL is a lousy language. My knowledge of C is more hearsay than hands on experience, so I will not say much about it. I do know some knowledgeable programmers who consider C to be a user hostile, counter intuitive language.

Only the millions (perhaps billions) of lines of existing code make it a viable language. That plus the huge number of COBOL trained programmers make it a language of choice for large companies. These are reasons to learn it, not reasons to like or consider it a good language.

The following omissions from COBOL make it inherently inferior to any language which provides them.

• COBOL does not provide Scope of Variables, which PL/1 and Algol provide in the Mainframe world. In the PC world, both VB and C provide this feature.
• COBOL does not allow you to declare variables within a Subroutine or Function. Even without scope of variables, this feature provides for better structure and makes a program easier to understand & debug.
• Both Algol and PL/1 allow Subroutines and Functions to be embedded within Subroutines and Functions. This provides a basis for well structured programming that even VB does not provide. In fairness to VB, this is not as important in the PC world due to the event driven nature of PC programs, which results in program structure different from either the PC DOS or the Mainframe worlds.

Ask your teacher to comment on the above omissions from COBOL. You can tell him/her that I have been programming since the early fifties, and have worked with assembly language for several mainframes, COBOL, Algol, PL/1, VB, and several languages nobody remembers anymore (EG: RUNCIBLE, ALTAC, and a few I would have to concentrate to remember, like Simscript).

Having said the above and before continuing, I still believe there are practical reasons for learning COBOL. While in school, I learned a lot that I did not like and which was inferior to other subjects I might have studied. A lot of the crap turned out to have practical value, and a lot (like history) turned out to be interesting to me later in life. Do not assume that I am arguing in favor of your not learning COBOL. It could turn out useful, even if it only provides some mental discipline or other general purpose intellectual value.

In farirness to your teacher: If you use some of my remarks, you should mention that I do not advocate not learning COBOL.

BTW: While I do not regret it, I suffered quite a bit for being a rebel in my scholastic days. Be ready to pay a price for not going with the flow.

Before it had some Algol & PL/1-like syntax added to it in the last 20-30 years, it was really a mess. From an esthetic point of view, it still smells. I think it still has a punched card orientation with optional line numbers that have no real purpose. I was a member of the original Codasyl Committee that designed the initial version of COBOL back in the late fifties & early sixties. Grace Harper (Hopper?) was the only one of us that really liked it, and I think she liked it because it had a lot to do with making her a big cheese in the early computer field. Do not get me wrong, she was a bright and competent person. The original version was terribly kludgy due to its being designed by a committee of about 25 members, some of whom were good coders who knew zilch about designing a language. Algol was a contemporary of early COBL and was considered a better language, so even from a historiical point of view, COBOL was not a big deal. If IBM had Algol and Burroughs had COBOL back in the sixties, nobody would even remember COBOL today.

A lot of good features were omitted for political reasons, or due to it being necessary to compromise, or because some dummies just did not understand about concepts like scope of variables and structured programming techniques. A committee of more than 5-10 people with equal authority cannot do a good job of designing anything. The effective IQ of a committee can be approximated by dividing the IQ of the smartest member by the number of members.

If a professional programmer had his choice of one language to be used for the rest of his career, only COBOL programmers would choose COBOL, and a lot of them would switch to something else. In the mainframe world, I doubt if any PL/1, Algol, or Pascal programmer would switch to COBOL. Nobody in their right mind would switch from VB to COBOL in the PC world, although some might switch from C.

I personally would choose Algol, with Pl/1 as second choice for Mainframes and probably VB for the PC world. I think a huge number of knowledgeable programmers would agree with me on these choices. I might consider Visual C for the PC World, not because I like the language, but because it seems to offer some advantages in speed and memory utilization. It also seems to have some deployment advantages.

I do not think the advantages of C are inherent in the language itself. I think they are the result of how VB & C have been implemented by Microsoft. A better implementation of VB might overcome any advantage C has now.

For ease of use and ease of learning, esthetics, and rapid development, I think VB is the easiest of all the powerful languages. It is esthically far superior to anything I have seen, although the PL/1 Select beats the VB Select Case hands down both in practicality and esthetics. The VB ElseIf is equivalent to the PL/1 Select in function, but its esthetics suck.

If Algol or PL/1 were put into a Visual environment for the PC, they would have to be any knowledgeable programmers language of choice. Nothing could help COBOL be anybody's language of choice. These languages (especially Algol) were inherently structured before anybody used the term, and could easily be enhanced to include features like classes, inheritance, encapsulation, et cetera.

Well, this is a little off-topic but I just thought I'd stick up for C and C++ a little. I have admittedly fairly recent and comparitively brief experience with programming (I started programming seriously, not just messing around, at 16, and I'm now 19) but having nearly completely migrated from VB to C/C++ over the past year I can say that I find it on the whole very sensible. I see far more logical behaviour in C that I do in VB. In VB it's sometimes difficult to see 'under the hood' and find out what's going on, where the bits and bytes are, but in C it's all easily to hand.

Having said all that, if I was developing any kind of office app I would probably favour VB (mainly I do games/graphics stuff so C suits me).

Perhaps a better implementation could improve performance significantly but VB is, by its very nature, abstracted to some degree compared to a lower level language like C. I don't really see how it could be any faster than MFC in C++, but I would certainly go with VB over MFC unless forced to do otherwise. In case you don't know (you said you weren't that familiar with C, or words to that effect), MFC stands for Microsoft Foundation Classes, and it's basically a way of using large, bulky classes and runtime DLLs to make Windows apps event-driven and easier/quicker to develop. I find it slightly silly really, since it's functionally like VB but with more complex syntax.

I get the impression that Delphi can, in some ways, give you the best of both worlds. Its speed and efficiency rivals that of C++, yet it is intended as a RAD language, with a nice clean interface like VB. I can't really comment on how easy to use it is since I haven't used it myself (I don't really like Pascal syntax) but many people seem to like it.

Many people seem to like it (delphi) and it is quite possible more powerful than VB but i would still choose VB over Delphi simply by the numbers of jobs there are in VB compared to Delphi.

I dont know why you dont like pascal syntax. I find it clear and consistant, and easy to read/write. Id take pascal syntax over cobol any day.

Well I agree that Pascal is better than Cobol, but so is just about anything. Pascal syntax seems a bit too heavy, it's not concise enough for my liking. I much prefer C syntax.

I prefer C-style syntax as well, so I prefer languages like C/C++, Java, Perl, etc.

It can be annoying how compact C++ (havt used C) is tho, coz it can make code hard to understand coz
A) There can be loads of commands all in one command
B) Its all symbols and stuff,eg ! instead of not so u end up with code that just looks a mess like

!(D++&F)==PQ&M*2 // (that probably means **** all so dont go posting messages saying it deost make sense coz i was just generalising!)

But it doesn't make sense :p (I don't either)

Hehe...it won't compile, but I understand your sentiment...

I love the :? operator for mind-destroying stuff :)

Heh, well I like that compact style :)

Parksie, do you still have that 12 Days Of Christmas code?

Quote:

---

Originally posted by Guv
Ksin: The comments I made in a previous post related to reasons for learning it, not reasons for liking it or considering it to be a good language. Compared to PL/1, Algol, VB, and a few others, COBOL is a lousy

---

It weren't me who started the thread honest Guv!
It were lord_dude ;)

Serious - very informative response - thanks.