ken@rochester.UUCP (Ipse dixit) (01/04/86)
References:
Info-Modula-2 Digest Saturday, 4 Jan 1986 Volume 1 : Issue 3
Today's Topics:
Alternative Source for Volition Compilers
More on Cambridge compiler
Std IO libraries
Ada (tm) vs. Modula-2
Impression of Logitech 2.0 system
Administrivia and freebie software
------------------------------------------------------------
Date: Sun, 15 Dec 85 21:14:34 est
From: ihnp4!inmet!tower@seismo.CSS.GOV
Subject: Alternative Source for Volition Compilers
To: info-modula-2
Springer Verlag was selling both the Logitech and Volition Modula-2
last time I checked. They are a computer science publishing firm
with offices in New York City and overseas. If you can't get their
address/phone number from (212) 555-1212, I'll see if I can dig it
up.
Len Tower
Intermetrics, Inc.
UUCP: {bellcore,ima,ihnp4}!inmet!tower
INTERNET: ima!inmet!tower@CCA-UNIX.ARPA
p
------------------------------
Date: 18 Dec 85 10:23:35 EST (Wed)
From: Ipse dixit <ken>
Subject: More on Cambridge compiler
To: info-modula-2
- ------- Forwarded Message
Date: Wed, 18 Dec 85 10:15:49 GMT
From: Peter Robinson <pr%computer-lab.cambridge.ac.uk@cs.ucl.ac.uk>
To: Info-Modula-2-request
Subject: Re: Info-Modula-2 Digest V1 #2
I noticed the comment from Arijit Das in Info-Modula-2 Digest V1 #2 and
thought that you might like some more recent news about the work with Modula-2
at Cambridge. I am afraid that rather a lot has happened since we distributed
the system described there, so this is going to be a long message...
A number of Modula-2 systems have been developed at the Cambridge University
Computer Laboratory since 1982. These all use the first passes of the Zurich
M2M compiler (albeit modified to remove the 16 bit dependencies), and so
accept the same, standard language. The run-time libraries have, however,
been written from scratch to our own design; but the libraries for input and
output described in the book Programming in Modula-2 can be implemented on top
of these.
The main system is for the DEC VAX under Berkeley Unix (or Ultrix). This uses
parts of the Zurich M2M compiler to produce a tree, from which binary is then
generated directly. The code is of a fairly high quality - comparable with c,
about 10% slower than Mike Powell's compiler from DEC WRL. This should not be
confused with the interim compiler distributed before 1984 which was slower by
a factor of about three. Our linker checks the module keys, and invokes the
standard Unix loader to build an executable program.
The Cambridge system includes an extensive program development system built
around the editor, emacs, the standard Unix make program, and our own symbolic
debugger, which can operate both interactively and post mortem.
The system is distributed on 1600 bpi magnetic tape in tar format, and
includes machine readable manual entries.
A second system is a cross-compiler for the Motorola 68000 which is written in
Modula-2 and operates under the Cambridge Modula-2 system for Berkeley
VAX/Unix. Again, the compiler uses the Zurich M2M compiler to produce M-code,
which is translated into assembly code or binary by a separate code-generator.
The code quality is quite good - comparable with c - but rather large.
This has been boot-strapped to run as a resident system on the 68000s running
Tripos as part of the Cambridge distributed system. The development tools
have been transferred as well (relatively easy, since they are written in
Modula-2) including a version of the debuger that operates remotely over the
network. We believe that other people have ported this system to 68000s
running Unix.
If you have some other 68000 Unix system, you will probably have to do some
work on the code-generator and the run-time library and perform the boot-strap
from scratch. To achieve this you must have access to another Modula-2 system
(which must support programs larger than 64k bytes), and have access to a tape
unit. We can supply the program sources on a 1600 bpi tar tape. The VAX/Unix
system would be an obvious environemnt in which to build and run the 68000
cross-compiler.
A further system has been developed in collaboration with a local company for
the National Semiconductor 32016 to run under their proprietary operating
system. At other times there have been compilers for the ICL/Three Rivers
Perq and the GEC 4090, but these have fallen into disrepair.
All these systems share the same run-time libraries (written substantially in
Modula-2), thus affording a high degree of portability. These provide a
system for input and output based on the use of streams which can be connected
to terminals, files display windows or whatever. A radical change was made to
this library early in 1985 when an additional stream parameter was added to
all the I/O library calls in place of the `currently selected streams' of the
older systems.
A further library implements a graphics package along the lines of the GKS
version 7.0 proposals, with drivers for a number of devices. The scheme for
graphical input differs radically from GKS by using a command interpreter for
interaction handling that provides independence from the actual tools
available.
These systems are distributed under the standard Zurich terms for a handling
charge of 100 pounds Sterling.
For copies of the licence agreement and full instructions, send your
snail-mail address to modula-2-request%UK.AC.Cam.CL@CS.UCL.AC.UK (or
modula-2-request%UK.AC.Cam.CL@UCL-CS.ARPA for Internet users who are not yet
using domain-based mail servers) or ...!mcvax!ukc!Cam.CL!modula-2-request.
- ------- End of Forwarded Message
------------------------------
Date: Wed, 18 Dec 85 08:18:33 pst
From: Vincent Broman <broman%bugs@nosc.ARPA>
Subject: std io libraries
To: info-modula-2
i have received from bob nagler two sets of definition modules plus commentary
for a standard io and system services library in modula-2. one is from the
british standards institution's modula-2 ad hoc library working group, a
distinguished set of m2 implementors, the other is a "cleaned up version"
proposed by bob nagler. they are both provisional, but worth examination.
these are available to arpanet users via anonymous ftp login at nosc-bugs.arpa
or bugs.nosc.mil in pub/modula/adhoc.lib/* and pub/modula/nagler.lib/* .
has anyone done implementations of these for either powell's or cambridge's
compiler on 4.2bsd unix on a vax?
Vincent Broman INTERNET: broman@nosc-bugs.arpa or bugs.nosc.mil
Analysis Branch, code 632 UUCP: {ihnp4,decvax,akgua,dcdwest,allegra,ucbvax}!
Naval Ocean Systems Center sdcsvax!noscvax!broman
San Diego, CA 92152 USA Phone: (619) 225-2365
------------------------------
Date: Wed, 18 Dec 85 12:20:11 PST
From: sun!orb!nagler@ucbvax.berkeley.edu (Rob Nagler)
Subject: Re: std io libraries
To: info-modula-2
Slight correction on Vincent Broman's note. It wasn't the British
Standards Institute's Adhoc Working Group, it is just an Adhoc Working
Group formed of Wirth's Collegues and some U.S. people. As of this
moment, the BSI is still working on the specification of what a
M2 library should look like.
Rob
------------------------------
Date: Sun 22 Dec 85 14:32:50-EST
From: Peter G. Trei <OC.TREI@CU20B.COLUMBIA.EDU>
Subject: Ada (tm) vs. Modula-2
To: info-ada@USC-ECLB.ARPA, info-pascal@BRL.ARPA, soft-eng@MIT-XX.ARPA,
info-modula-2
The current issue of ACM SIGPLAN Notices carries an article,
"ADA(*) VS. MODULA-2: A VIEW FROM THE TRENCHES" by Richard Bielak,
contrasting the two as tools for real-world applications programming.
Ritchie is a colleague of mine who lacks net-access, but I will
forward any comments. I am not on all the lists I am mailing this to,
so please send your comments directly to me. The article is in SIGPLAN
Notices, Vol. 20, #12, (Dec 85), pp 13-17.
Peter Trei
oc.trei@cu20b.arpa
(*) ADA is a registered trademark of the U.S. Government.
- -------
------------------------------
Date: Thu, 2 Jan 86 14:22:52 CST
From: Dan Frank <frank@db.wisc.edu>
Subject: Impression of Logitech 2.0 system
To: info-modula-2
Well, I've installed my version 2.0 of the Logitech compiler and
new syntax-checking editor, and I am impressed. Execution speed is
significantly improved (of the compiler and linker, that is), and
while the editor is not blindingly fast, it is pleasant to use, with
a fast syntax checker, syntax sensitive indentation control (!), a
really slick template facility that allows you to specify where the
cursor should come to rest after the editor inserts your prototype
statement block. The editor supports tiled multiple file windows,
but doesn't burden you with an outline around the screen.
Like turbo, you can compile, link, and execute from within the
editor. If you have compilation errors, the editor will read the
listing file and allow you to step through the file error by error,
with the offending line and error message displayed in a window
near your source file line.
Note that this is not a syntax-directed editor, or an incremental
compiler, or anything of the sort. It just has a key you push that
casuses the text buffer to be scanned and parsed. Real fast, and
quite useful, since most compilation errors are syntax errors, anyway.
If you get the professional package, you will get about nine diskettes.
If you put up the whole thing, it may eat about three megabytes. This
is not for the faint hearted or short of medium. The manual, by the
way, has been completely rewritten and much improved. It is still
oriented toward the software systems engineer, though; I've always
liked the fact that Logitech flatters rather than insults my intelligence,
but it may intimidate the first time user.
All in all, I highly recommend the upgrade for current users, and
the whole system, at its new, very reduced price, for anyone serious
about programming on the PC. I've been doing some consulting in C
over vacation, and it really brings home to me why I like Modula-2
so much!
-- Dan
------------------------------
Date: 03 Jan 86 21:30:58 EST (Fri)
From: Ipse dixit <ken>
Subject: Administrivia and freebie software
To: info-modula-2
This mailing list has seen little activity over the last 3 weeks due to
the holidays. I would like to take the opportunity to wish everyone a
Happy New Year.
We now have an archive for public domain Modula-2 software. It is at
SIMTEL20.ARPA in area PD:<MODULA-2>. If you have Modula-2 software to
share, here is your chance! I made up a list of possible guidelines for
discussion. I'm new to this coordinating business so if there are any
older hands out there who see something I have missed or have better
suggestions, I'd be glad to have your comments.
BEGIN proposed_guidelines
1. Submitted sources should be free of any proprietary restrictions.
Please indicate which category your submission falls into, e.g.
games/educational, utilities, languages, etc. Remember to document any
compiler/system dependencies.
2. Generally, programs should be submitted in source form. If it is
felt that there will be demand from people without access to a
compiler, then binaries may also be submitted with the source. Binary
only submissions are discouraged.
3. Multiple files should be bundled together before submission to
reduce confusion. It should be possible to unbundle with nothing more
than a text editor. For example the Usenet "shar" format is
acceptable. Even concatenating the files with prominent separating
lines is OK. The first file in the bundle should be a README file,
followed by documentation, if any, then the programs. If the combined
size of the bundle exceeds 64 Kbytes, then it should be split into
parts. This is to avoid problems with mailer limitations.
4. Binaries should be archived in some standard encoding, to be
determined. I would like to propose some simple standard, like
uuencode/uudecode. Many of the list participants do not have FTP access
and must rely on mail to receive programs.
5. To avoid innundating the Keith Petersen at SIMTEL20 with frequent
updates, sources should be released for a trial period of 2 months, or
until a new revision is announced, whichever happens sooner. At the
beginning of the trial period the sources are deposited in the
archives. During the trial period, the sources may be obtained from
the archive, or the most recent version from the submitter, or other
redistributors. Bug reports and fixes should be sent to the submitter.
For this purpose, the README file should contain a reply address. After
the trial period, the updated sources should be sent to the archive to
replace any superseeded versions.
6. The above are simply guidelines. Common sense should be the final
arbiter.
END proposed_guidelines
If you would like to submit your favourite brainchild or have queries,
mail me and I will get things going!
To start the ball rolling, I am offering my "new, improved"
prettyprinter again. Several bugs have been fixed since I last
announced it in November. Mail me if you want a copy. I am interested
in more bug reports, suggestions for improvement, especially in the
area of modularity, portability problems, or what have you.
Ken
- --
USnail: Dept. of Computer Science, U of Rochester, Rochester 14627.
UUCP: ..!{allegra,decvax,seismo}!rochester!ken ARPA: ken@rochester.arpa
Phone: (716) 275-2569 (office), (716) 262-5298 (home) Voice: Ken!
------------------------------
End of Info-Modula-2 Digest
***************************
--
UUCP: ..!{allegra,decvax,seismo}!rochester!ken ARPA: ken@rochester.arpa
USnail: Dept. of Comp. Sci., U. of Rochester, NY 14627. Voice: Ken!