I wouldn't wish a COBOL career on my worst enemy. What a god-awful mess. And yes, I know how and why it came to be. And all about legacy code bases and the high cost of migration. I was forced to take cobol back in 1978, never attended a lecture, handed in all the assignments, wrote the final, passed the course, and promptly forgot the whole thing. Still, what a god-awful mess :-)

From: Ethan Dicks Sent: Tuesday, February 17, 2015 11:03 AM [snip] As late as 1990, I was still telling student employees at LOTS that they ought to learn COBOL and JCL (MVS and DOS/VSE) as well as the cool languages. That way they were always assured of finding work while they looked for a job. I don't think any of them believed me... I learned FORTRAN IV (1401) and PL/1 and COBOL (360/40 under DOS) my last semester of high school, then picked up 360 Assembler as well as improving all three over the following summer. I only used Fortran in anger for a short while in the late 1990s, but I made good money with COBOL and PL/1 through college, grad school, and for a while beyond, until I moved into systems programming (SVS/MVS and TOPS-20). In the mean time, I wrote a Pascal compiler for the DEC-20, and began my lifelong love affair with Lisp. My *only* forays into BASIC were to read the Tiny Pascal compiler articles in BYTE, and the source for the Star Trek game that ran under Wylbur on the UChicago Wylbur system. I don't think COBOL has hurt me at all. In fact, I considered brushing it off before I landed the current position, and I've had occasion to write some for real production use at the museum. Rich Rich Alderson Vintage Computing Sr. Systems Engineer Living Computer Museum 2245 1st Avenue S Seattle, WA 98134 mailto:RichA at LivingComputerMuseum.org http://www.LivingComputerMuseum.org/

On Tue, 17 Feb 2015, Fred Cisin wrote: Djikstra is a bloviating jackass. g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://scarlet.deltasoft.com - Get it _today_!

On Tue, 17 Feb 2015, ben wrote: End of February 1979, Microsoft released FORTRAN for the TRS80. They had announced it months before, but it got delayed. I got the first few copies on my way home from the eclipse (had Bob Wallace meet me in the Seattle airport, otherwise, it would have been delayed further) Microsoft FORTRAN for the PC was unimpressive. Integer sieve of Erastothanes ran slower than interpreted BASIC! -- Grumpy Ol' Fred cisin at xenosoft.com

On Tue, 17 Feb 2015, Fred Cisin wrote: What version was that? In 1991 I was using Microsoft FORTRAN on the PC for ground water modelling. The floating point matrix math was much faster than anything I could do in BASIC by several orders of magnitude. I was a poor engineering student so this was my on my old Tandy 1000A with a 286 accelerator and 287 math chip. Kelly

I was offering a perspective on their REAL type 9 years later. I hope they made some improvements in that time. It was quite fun to use and then capture the output into a file that I then parsed into influence graphs and solute transport graphs based on the aquifer parameters.

So long as it wasn't DEC Fortran it was portable.....

Zitat von Eric Smith <spacewar at gmail.com>: Finally somebody who liked Pascal ;-) I enojoyed playing with Pascal/Modula etc. However, paid jobs were FORTRAN & assembler back then

On Tue, Feb 17, 2015 at 11:33:58PM -0700, ben wrote: [...] So, where does Perl fit in? :)

At 01:33 AM 2/18/2015, Ben wrote: And that was the original purpose of Pascal. Niklaus Wirth initially designed Pascal as a procedural language to teach structured programming. In the CP/M days I experimented with other languages (I had Mix C, Turbo Modula-2, and others that I got pretty cheaply but never did much with) but ended up sticking with Turbo Pascal optimized with some assembly language routines. In high school in the late 1960s, at the MIT High School Studies program, I looked at a number of languages but ended up studying and working mostly with FORTRAN and ALGOL because my primary interest was in scientific computing. I wrote some programs in SNOBOL because string manipulation intrigued me. I took one look at a COBOL manual but realized that I had no interest in business applications. Dale H. Cook, Roanoke/Lynchburg, VA Osborne 1 / Kaypro 4-84 / Kaypro 1 / Amstrad PPC-640 http://plymouthcolony.net/starcity/radios/index.html

It was thus said that the Great Chuck Guzis once stated: Um ... for i = 1 , 23 do para_1() para_2() para_3() para_4() end para_1() para_2() para_3() para_4() para_1() para_2() repeat para_2() para_3() index_1 -= 3 until flag_1 > 0 Am I missing someting? -spc (I might have the logic messed up on that last one, not knowing COBOL)

From: Sean Conner Sent: Wednesday, February 18, 2015 8:56 AM ^^^^^^^^^ ^^^^^ Syntax error. The syntax of the PERFORM verb in this case is PERFORM paragraph-1 [{THRU,THROUGH} paragraph-2] VARYING {identifier-1, index-name-1} FROM {identifier-2, index-name-2, literal-1} BY {identifier-3, index-name-3, literal-2} UNTIL condition-1 [AFTER VARYING {identifier-4, index-name-4} FROM {identifier-5 index-name-5 literal-3} BY {identifier-6, index-name-6, literal-4} UNTIL condition-2 [AFTER VARYING {identifier-7, index-name-7} FROM {identifier-8 index-name-8 literal-5} BY {identifier-9, index-name-9, literal-6} UNTIL condition-3]] . Braces indicate alternatives, brackets indicate optional portions of the statement. An AFTER VARYING clause can only occur in a 2nd or 3rd position following the 1st VARYING clause. There is no "WITH xxxx" in the 1968 or 1974 CODASYL standard; did that come into the language in 1985 or later? Never mind, I just checked the IBM COBOL reference manual, based on the 1985 standard, and do not see it there, either. index_1 = 10 ^^^^^^^^^^^^ Sean forgot to initialize index_1 to 10 before beginning the repeat ... until loop. Rich Rich Alderson Vintage Computing Sr. Systems Engineer Living Computer Museum 2245 1st Avenue S Seattle, WA 98134 mailto:RichA at LivingComputerMuseum.org http://www.LivingComputerMuseum.org/

On Wed, Feb 18, 2015 at 9:55 AM, Sean Conner <spc at conman.org> wrote: No. In the COBOL example PARA-1 though PARA-4 are consecutive paragraphs, not functions that are terminated by a return (explicit or implicit). If you want to think of it as a function, the paragraphs are all part of a single function, but both the entry and return points of the function are variable, specified by the caller! No one claimed that it wasn't possible to achieve the equivalent effect of such a program using normal constructs of other languages; it would just have to be structured differently.

On Wed, Feb 18, 2015 at 6:52 PM, Rich Alderson <RichA at livingcomputermuseum.org> wrote: Point taken.

On 18 February 2015 at 22:41, Toby Thain <toby at telegraphics.com.au> wrote: hours coding with it. I wouldn't chose to use it now, but back in the day it was the best development system for CP/M and MS-DOS. -- Yoyodyne Propulsion Systems: "The Future Begins Tomorrow" Visit us at: http://www.yoyodyne-propulsion.net -------- "When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him." -- Jonathan Swift

I was introduced to Pascal in a high-school AP computer science course ca. 1986. It was a real eye-opener, and seemed quite powerful after only using BASIC and the occasional chunks of borrowed 6502 ASM. Personally, I thought it was an excellent teaching language - at least for HS students in the late 1980s. I really enjoyed learning to use it - though as someone who'd had almost ten years' time with BASIC, I found the emphasis on recursive (sub-)routines to be a bit difficult to grasp. Something about it just seemed... kludge-y. But functions were a very welcome addition.. Of course, I've forgotten it all! On Wed, Feb 18, 2015 at 5:05 PM, Ethan Dicks <ethan.dicks at gmail.com> wrote:

On 2/18/15 6:03 PM, Jon Elson wrote: Back in the late 70's and early 80's I was busily writing a number of different Pascal compilers including one that I wrote while I was at IBM (I think that was number 6). Bill Maddox had actually unearthed one of my middle Pascal compilers that I wrote while an undergrad at CMU along with all of the documentation. The claim to fame of that compiler was the development/documentation of the stack machine (and associated tools, including an interpreter) that was the target of the compiler. All of it was written in Pascal. A later version (and significantly restructured/rewritten) was to have an interpreted stack machine that would be run on C.MMP. That version is at this point lost to the mists of time. :-( BTW, all 6 of the Pascal compilers that I wrote were all written in Pascal, used recursive decent parsers and were self hosted (that is the compiler when working would compile itself) which was a great test of the entire tool chain. ;-) TTFN - Guy

Aw, come on, Dave. So long as it wasn't *any* manufacturer's Fortran, it was portable. The first textbook I used, _FORTRAN IV Primer_ by Elliott I. Organick, was filled with tables for all the various places that implementations differed among manufacturers. (That was where, in fact, I learned that there *were* manufacturers other than IBM.) No DEC systems were on the list. DEC was very careful to mark out where they had extended the language from the 1966 standard (and later from the 1977 standard, which included most of the DEC extensions to 1966). If you restricted yourself to the standard, Fortran was very portable, and that stood it in good stead for decades. Good programmers put any of the manufacturer specific code in subroutines which were easily replaced. Rich Rich Alderson Vintage Computing Sr. Systems Engineer Living Computer Museum 2245 1st Avenue S Seattle, WA 98134 mailto:RichA at LivingComputerMuseum.org http://www.LivingComputerMuseum.org/

On 02/17/2015 02:39 PM, ben wrote: Actually, it was entirely possible to write portable programs in any FORTRAN IV or later system. Note that this included *DECIMAL* machines as well as non-ASCII, non-byte addressable machines. As far as I know, C doesn't exist for decimal machines. Somewhere, I've got a Pascal compiler written in FORTRAN. It was very usual on non-number-crunching portable programs to see: INTEGER ICSET(80) COMMON /CHARSET/ ICSET READ 100, ICSET 100 FORMAT( 80A1) Or something similar. The purpose is to get the basic character set defined that you'll be using, so the first input card would have the letters and numbers of the character set in a particular order. --Chuck

College class project COBOL was NOTHING like working on years old ( or decades old O_o ) production COBOL. It usually entailed going on an archeological dig through source, trying to discover business rules that nobody left on the devl team could remember. Also re-discovering why some pieces were "architected" the way they were. Those brave kids will earn any high $$$$ with blood, sweat, and many many tears. TKilling On Tue, Feb 17, 2015 at 2:27 PM, Mike Stein <mhs.stein at gmail.com> wrote:

On 02/17/2015 10:54 AM, Roe Peterson wrote: Try *maintaining* the COBOL compiler for a couple of years. Mostly, you sit with a pile of PSRs (program status/trouble reports) and the CODASYL report, quoting chapter and verse to the submitter. Writing programs in the stuff was actually fun after that... ...and then try maintaining a custom version of COBOL where you had chains of communicating COBOL modules spread out over several machines--and where a module could manipulate the data of another that wasn't even resident in *any* machine's memory... Good times--I've tried to forget about them. I really have. --Chuck