Who's into FORTRAN?

[CP/M User]

Wikipedia experience

Wikipedia experience

carlsson wrote:

Yeah, an encyclopedia written by the people is a dual edged sword. Often it will end up with actual, helpful information, but in matters of history or as in this case people, it could become ugly. However, I suggest you write your own accurate story, point out the inaccuracies in tabloid press and subsequently Wikipedia. Find help from someone about search engine optimization to make your page the first one to appear on a web search.

Anyway, it is refreshing to see that Wikipedia is about posting stuff that can be verified, not neccessarily the truth. ;-) A commercially edited encyclopedia would probably work in the other way around, trying to verify all sources to find the truth and nothing but the truth.

Naturally I feel into their trap by including a link to my website and when my old site went belly up, I updated it on their Wikipedia and got busted! I guess I should have read their terms and conditions to know what you can and cannot do. Though they had this article and there were several external links to simular pages which shouldn't have been there either, those pages got removed (thanks to me!! ). What I don't get is the Turbo Pascal 3.0 Compiler and Code Generation Internals remains even though it's not really Pascal related - even though it goes through discussions on little bits of pascal code, the site I felt is about disassembling Turbo Pascal 3 and going through the program and Library section.

But it doesn't worry me!

 

[Chuck(G)]

Nothing like a 4 year old thread.

I've written quite a bit of FORTRAN (or Fortran, as it's now called), been a principal in a firm writing a new optimizing, vectorizing compiler and an alternate rep to X3J3 (Fortran 90, which was called at the time FORTRAN 8x).

One thing that Fortran (or FORTRAN) is hugely suited to is number-crunching. It's easier for an optimizer to generate really tight code from Fortran than almost any other language. Several PC products were originally written in Fortran, as it was the most portable language at the time.

Early on when microcomputers were first making their appearance, development systems had not yet evolved sufficiently, so cross-compilers and cross-assemblers were used. Almost to a product, they were written in Fortran. Every mini and mainframe had Fortran on it and it was possible to write a completely portable cross-assembler or cross-compiler. That was a lot tougher back then--not all systems in the 70s had 8 bit bytes, used ASCII or EBCDIC or even used twos complement arithmetic. You wrote your programs on punched cards mostly.

Fortran could compile (not interpret) on systems with as little as 4K of memory and no disk or tape storage. I could do it on a decimal (not binary) IBM 1620 with 20,000 digits of memory and still have lots of memory left over.

There's a very small USA BASIC FORTRAN from the 1960s with very few statement types, if you want something that's easy to learn.

 

[CP/M User]

I don't know what I was doing wrong when I was using Digital Research's own variation of Fortran-77. I was using it in CP/M-86 v1.1 mind you and the Compiled (".CMD") File was over 64Kb in size and that was after I'd optimized the time, the unoptimized file was even bigger, cannot remember how big it was now unfortunately! Having said that I used a Free Fortran-77 compiler I downloaded years ago from Simtel for DOS which didn't have the same problems when compared to the DR version, maybe DR had to throw in this huge library file, just to get it to work?

 

[Chuck(G)]

Don't know--Microsoft's F80 was pretty good on the x80 platforms, though. I used it for quite a while--even wrote a filesystem conversion (ISIS-II to CP/M) utility in it.

 

[Anonymous Freak]

My experience with FORTRAN was as a required course in my Mechanical Engineering coursework. The professor was discussing the possibility of moving to FORTRAN 90 from FORTRAN 77. This was in 1996. What I remember more than anything was going through all of the assignments in the first week, and having the coding done; then spending the rest of the term adding chrome, like progress bars. Do you know how hard it is to code a progress bar in raw FORTRAN?

The other thing was that it was a class of ~40, in a lab with 50 computers; but a networked compiler that only had licenses for 5. And often times, the licensing "system" would get confused and not release a license. I dug in, and discovered that the "licensing system" was really just a DOS batch file that created five temp files, and if all five existed when you launched the batch file, it spat out an error. Of course, it was SUPPOSED to delete the highest-numbered temp file when your compiler exited, but often didn't; which "hung" the licensing.

So I just went around it and ran the raw command, rather than using the "license checking" batch file.

Before figuring that out, I owned an HP 200LX, and found that I could compile faster on there than on the computer lab systems, solely because even though the 200LX was slow as dirt, it was faster than waiting for a license.