jbn@wdl1.UUCP (John B. Nagle) (04/11/84)
I am interested in using Modula II on an IBM PC or equivalent. Questions: - Who sells a compiler? - Is it reliable w.r.t. producing correct code? - Is pointer/subrange/array run-time checking available? - Is it fast compared to (say) Lattice C? - Can it support large arrays (>64K) and pointers beyond 64K? Nagle@SCORE
todd@islenet.UUCP (04/19/84)
I just bought the $40 Modula-2 compiler from Modula Research Institute
(phone 801-375-7402 for order forms & licensing terms). Unfortunately,
I have not had enough time to really make some tests (or even read the
entire manual for that matter!). However, I can say the following:
1. It works as advertised and expected
2. It is the only current Modula-2 implementation for the ibm-pc
that supports the type 'real' without an 8087 installed
3. It comes with enough info to write assembly routines for
4. It comes with its own linker
5. Source code for the compiler is available for $160
6. It compiles very slowly (4 pass compiler)
7. It compiles to M-Code (not native)
8. Due to '7' it executes reasonably quickly (much faster
than interpreted BASIC or BASICA but not nearly as fast
as Lattice/Microsoft C (but then what does on the pc?)
9. M2 from MRI at $40 is definitely worth getting for
learning purposes (I don't have any other M2 compiler
available to me). However, I would not use it to write
production stuff.
If you want a good fast compiler, take a look at Turbo-Pascal. It
has enough extensions to make it very usable too (case w/else, etc.).
Todd Ogasawara -- University of Hawaii -- Dept. of Psychology
{ dual, vortex, uhpgvax }!islenet!toddtapia@uiucdcs.UUCP (04/23/84)
#R:wdl1:-19200:uiucdcs:24700043:000:1508 uiucdcs!tapia Apr 22 19:04:00 1984 There are 3 companies that sell Modula-2 compilers. 1) Logitech, a subsidiary of a European firm. Palo Alto, CA (415)326-3885, Approx. price $500 2) Volition Systems P.O. Box 1236 Del Mar, CA 92014 (619)481-2286 (no idead about price) 3) Diser Inc. (no additional info about this company) For references you might look in: BYTE December 83 pp.59-64, August pp. 324-327. I have learned about these firms because I'm trying to transport a Modula-2 compiler that runs on Lilith (the original Modula machine designed and built by N. Wirth's group) to an IMB PC. I suspect that Logitech did something similar already, i.e., they took the Lilith compiler that runs on a PDP-11 (on an M-code interpreter) and modified its forth pass to generate 8086 code. At the moment I am working with a version that runs on Unix in VAX compatibilty mode. So far I have had no problems with it. The documentation I have, however, is poor. This version, which is not very different from the one above since it was writen by the same people, has the capability to check for subrange and array run-time errors. I have no comments to make on its speed, except for the fact that it isn't any slower than any other software running on our system. On arrays > 64K, I think, the problem is not a language problem (Modula, or any other) but a limitation of the 8086 chip, i.e., you have to work with 64K segments. Cris Tapia Univ. of Illinois, Champaign-Urbana Dept. of Computer Science