Slashdot Mirror

← Back to Stories (view on slashdot.org)

COBOL IDE, Compiler for Linux

Posted by michael on from the new-tricks-for-old-dogs dept.

Grizzly writes "Okay, go ahead and laugh. But there's a lot of COBOL out there, and that has kept a lot of businesses, especially outside the United States, from even looking at Linux. Linux and Main has a story on KOBOL and how it might make bringing those COBOL apps over to Linux possible."

26 comments

  1. 1st post by Anonymous Coward · · Score: 0

    1st post done.

  2. BTW, what "cob2c" compiler is this? by Anonymous Coward · · Score: 0

    I'm curious what "cob2c" COBOL to GCC compiler they are using? Is it an Open Source one? Or is it a licensed compiler from someone else -- possibly modified for GCC? Version 0.9.7 -- couldn't find a project/product with that version.

    1. Re:BTW, what "cob2c" compiler is this? by Arrgh · · Score: 2

      This identical question has already been answered in the original article.

  3. TinyCOBOL ? by Anonymous Coward · · Score: 1, Informative

    Are you aware of the TinyCOBOL project? They are at version 0.58 and producing a true COBOL compiler, no translation to C first. They are targetting COBOL 85 first with an eventual production of a compiler with the most current standard available.

    The project is hosted on SF.net at http://tiny-cobol.sf.net/, check it out for more details.

  4. back to the future by Pauly · · Score: 0, Flamebait

    "...it might make bringing those COBOL apps over to Linux possible..."

    Please don't.

    Programs written for mainframes suck no matter where you run them. Paying people to sit around to watch batch jobs run is stupid no matter whether they're watching a 3270 terminal or an xterm.

    1. Re:back to the future by Gaetano · · Score: 3, Insightful

      It sucks but when you are a medium sized company with huge amounts of Cobol applications running on a mainframe which you are paying huge amounts of money for you have to do something.

      Those Cobol applications are things that my transportation agency count on to do buisness, like the payroll system, and the parts maintence system for all the busses.

      We are looking at $450,000 to renew the lease on the mainframe in two years, or we can port the applications to our Unix environment that we already have and tell IBM what they can do with their mainframe.

      Its much easier and more cost effective to bring the code over nearly as-is versus migrating to a new language entirely.

    2. Re:back to the future by Pauly · · Score: 2

      Its much easier and more cost effective to bring the code over nearly as-is versus migrating to a new language entirely.
      If it really is, then bravo. You've made the right decision. In my experience however, most people making this decision haven't taken into account the complete cost of these applications. More than the absurb hardware maintenance costs, there is an elevated cost for running batch-oriented applications. How many times a week does a human have to intervene overnight when a job fails? How much does that human cost you per hour in salary, benefits, office space, etc?

      My point is that mainframes and the applications designed for them were built under the assumption that an hour of human time is orders of magnitude cheaper than an hour of computer time. Now, especially with high power to cost ratio systems like Linux/x86, that ratio has inverted. It's now much cheaper to let computers do the repetitive work and send the humans home. You might be able to save enough to justify spending nearly half a million dollars on some quality people to rewrite those applications for you.

    3. Re:back to the future by Gaetano · · Score: 1

      I agree with you in this. I don't think most people who are looking for Cobol solutions are looking for them to develop new applications. Like us they are trying to lower costs, while providing a migration path to other systems.

      For example, after our Cobol programs are ported to the unix system on Oracle we will be able to gradually replace parts of the application with another application that accesses the Oracle database (maintaining the buisness rules in the database) where both systems can operate in congunction untill we have perged all our Cobol code.

      This takes time. Maybe by the time we are done doing it the language that the replacement code was written in will me legacy and need to be replaced. :) But hopefully it will be done before our cobol programmers have retired.

      But if we port directly to another platform with another language ourselfs, or if we are porting to another software package, we are looking at a longer term solution where both systems will have to be maintained durring the migration (and thus more mainframe leasing costs) than we would be looking at if we just move the Cobol applications over to the Unix environment and port them from there.

    4. Re:back to the future by hughk · · Score: 2
      We have 30+MB of COBOL source code for one application, the C is less than 10MB. The applications runs on an OpenVMS Cluster and OpenVMS middle-layer systems. It runs like hot-shit off teflon shovels. The systems runs in two modes, "ONLINE" doing a few million transactions per day and batch, where is does some cleanup work at the end of the day. Don't even *think* of doing this on a PC, the I/O bandwidth sucks big time and the OS design lacks a lot (even Linux). Sure we pay people to look after things, but we only need two shifts.

      We have a front-end implemented mostly in C and Java running under Win2K and Solaris. The GUI is just Java and it sucks big time.

      --
      See my journal, I write things there
  5. PerCobol LegacyJ by Gaetano · · Score: 3, Informative

    Another very attractive cobol tool that runs on linux (and windows, and Solaris, and so on) is PerCobol from LegacyJ.As the name would indicate its a java solution.

    http://www.legacyj.com/lgcyj_perc1.html

    My agency has had a developer trying to get Merants Cobol working on a Solaris system for some time now and hasn't had any luck getting it to connect to Oracle. All the instructions we could find refer to DB2, even those where incomplete. Merant was not willing to support us since we purchased the software directly from Sun.

    I couldn't find any instructions on getting Kobol to connect to Orable either, but its still a pretty new product so I will be revisiting it later. But with PerCobol it took about 1 hour to figure it out. (this is including the 30 minutes it took to download the 100mb evaluation copy).

    The cost of PerCobol is MUCH higher than Kobol's $39.95. Its more like 5000 dollars per developer (if you are deploying to a Solaris server which we are). But it still looks really good compared to what we have experianced with Merant and we can redeploy the mainframe applications to java applettes on our intranet with PerCobol with little trouble.

  6. Well... by vegetablespork · · Score: 3, Funny
    ...if obsolescence implies security, then your apps won't get much more locked down than with this!

    --

    Call (206) 338-5780 COLLECT for information about a genuine BA, BS, MA, MS, MBA, or Ph.D.

    1. Re:Well... by einhverfr · · Score: 2

      Cool Now YOU can keep your legacy, outdated programs when you migrate to state of the art systems! Wait, how did my insurance plan expire in 1902? It was supposed to do so in 2002 ;)

      --

      LedgerSMB: Open source Accounting/ERP
  7. Don't disrespect COBOL too much by PD · · Score: 5, Insightful

    Or at least don't disrespect the programmers. Someday when we're all creaky it'll be hard to find a job doing C++ or Java or Perl or C#. Sure, we know those are fine languages for many things, and we've got 40 years of experience using them. Sure we are still sharp and can code rings around almost anybody. We even probably still feel young inside and if we're lucky only need the regular strength viagra. (EXTRA strength viagra is for OLD people).

    But there it is on the resume - you did C++ programming 25 years ago, and in the eyes of the young'ins, that makes you unemployable. I've worked with enough people with many decades of experience to know that these people can be the sharpest ones in an organization, even if their experience includes COBOL.

    I know this is offtopic, but I can predict what some of the comments are going to say.

    --
    If tits were wings it'd be flying around.
    1. Re:Don't disrespect COBOL too much by Anonymous Coward · · Score: 0

      And it not just the creaky folks who have COBOL on their resume. I'm no longer a spring chicken, but I'm not elderly, either... and my professional career started with hacking COBOL for the U.S. military. The stuff's still lurking around out there....

  8. good for linux by larry+bagina · · Score: 4, Insightful
    I'm sure a lot of people will say "Cobol is dead. Who cares".


    While it's true that cobol is often associated with legacy programs, cobol programs are a niche market, one that linux could exploit (recompile your cobol app for linux and save!). So far, linux growth has mostly been at the expense of other unix rather than cutting into the desktop (windos) market. I'm not suggesting that linux or x86 hardware can compete with mainframes, but a modern x86 box running linux could replace a mainframe for some cobol apps.


    Additionally, scary as it sounds, ne wCobol development is still being done. A company I used to work for sold (and still sells) expensive bank software written in cobol and using Tandem SQL.


    If it gets the job done, it gets the job done. That applies to linux and cobol.

    --
    Do you even lift?

    These aren't the 'roids you're looking for.

    1. Re:good for linux by Anonymous Coward · · Score: 0

      Don't forget RPG while your at it. I wonder how many places went from IBM Model36-> Model38 -> AS400 beacuse all of there code was writen on in RPG.

  9. What these compilers miss by MrBoring · · Score: 5, Insightful

    Unfortunately a COBOL compiler isn't enough. Other applications which are specific to mainframes would also need to be implemented. Here's two examples:

    * VSAM -- IBM's keyed file format. It's simplistic in concept, but manages keyed files very well. A relational DB might be better, but you'd still have to convert the code.

    * IMS -- IBM's hierarchical database. The syntax isn't declarative like SQL, but much more intent on moving a pointer throughout the database.

    * Other OS calls.

    1. Re:What these compilers miss by cowdog · · Score: 1

      Indexed files are supported in Kobol. Not exactly VSAM, but close enough for most applications.

      As for IMS, there is a command level preprocessor available for it. It is just rarely used. In fact, it is the same preprocessor that is used for command level CICS.

  10. PowerCOBOL by Anonymous Coward · · Score: 1, Informative

    I was turned off to COBOL in the beginning, until I tried Fujitsu PowerCOBOL for Win32. Programming with it is much easier than I imagined it to be. I made an interesting GUI in a couple of hours. The code is pretty readable too. I really think it's a myth that COBOL is dead and legacy only. The recent COBOL book from Sams was what got me started, and came with this compiler on CD.

    It's not perfect however, there are some annoying bits about the IDE, at least in the demo that I tried (based on an older release). I had to close the editing window before I could compile. It becomes annoying opening and closing the same window during tests. Hopefully this issue is solved in newer releases.

    I'm pleased that there will be KOBOL for Linux. I hope it measures up to PowerCOBOL eventually.

  11. You're missing the most important thing by Garg · · Score: 4, Interesting

    Women.

    I've been in computers 20+ years, and worked with all sorts of different platforms and languages. For some reason, there are more women who program in COBOL than in any other language.

    You guys in the ultra-geeky C++ shops, look around. What's the ratio of males to females? In most COBOL shops, it's equitable. I've even worked in places where there were more women than men.

    I've recently interviewed for a job elsewhere in my organization where I'd be swapping mostly COBOL and a little Java for full-time Java, SOAP, etc. Sounds like a much cooler job... but this group is off in another building, and there's only one woman. Don't get me wrong, I'm happily married... but I like having women around.

    So what if all the women will look like Grace Hopper soon? I'll look like Hume Cronyn.

    Programming needs women. Go COBOL!

    Garg

    --
    Garg
    Alumnus, Xavier's School for Gifted Youngsters
    1. Re:You're missing the most important thing by jayayeem · · Score: 1

      I've had the same experience, and I have a theory about it. Cobol programmers, especially the long timers, are by and large not CS geeks. They are people whol came from the business side and recieved corporate training. C++ (and most other languages) geeks tend to be people that studied CS at StateTech.edu.

      Therefore Cobol programmers represent the heterogeneous make up of the business world, while other computer disciplines represent the nearly all male world of University CS classes.

      --
      I metamoderate, therefore I am
  12. Listen you whippersnappers by Cro+Magnon · · Score: 1

    COBOL is a great language for business/financial applications. I wouldn't write an OS in it, but I wouldn't recommend C for banking software either.

    --
    Slow down, cowboy! It has been 4 hours since you last posted. You must wait another few hours.
    1. Re:Listen you whippersnappers by MrCreosote · · Score: 1

      On the other hand, I wrote a couple of programs to generate COBOL data-entry and report programs, in COBOL.

      --
      MrCreosote Meow!Thump!Meow!Thump!Meow!Thump! "You're right! There isn't enough room to swing a cat in here!"
  13. Way OT, but what the hell... by Amazing+Quantum+Man · · Score: 1

    Is OpenVMS really open? How about an x86 port of VMS?

    --
    Fascism starts when the efficiency of the government becomes more important than the rights of the people.
  14. Reinventing Micro Focus? by broody · · Score: 1

    Micro Focus is going to eat the Kompany's lunch on this one. Mainframers know Micro Focus like C++ developers know Rogue Wave. Their Object COBOL developer suite has been here longer from a company Mainframers know and trust.

    I have nothing against the Kompany, in fact I like them quite a bit, but this seems like a stupid move.

    --
    ~~ What's stopping you?
  15. Ridiculous by krashish · · Score: 0

    You people have got to be kidding! Anything done in COBOL can be re-written in C 1000 times more efficiently, quicker, and easier to understand for the next (younger) guy in line for your job.

    Give me a break, COBOL, if this isn't satire, I don't know what is!