The Fred Fish Collection 1.5/ffcollection-1-5-1992-11.iso/ff_progs/prog_oth/comal.lzh/COMAL/INFOTEXT/HISTORY

Text File | 1991-08-16 | 38.5 KB | 985 lines

* Are you going to be moving around a lot during
your talk, or will you be fairly stationary?

What would you prefer?

* Well, I don't mind.

I promise you that I will no do Danish Folk
Dancing.

* Oh.

That would be a disaster. We would probably be
thrown out of NATO afterwards.

* Anytime you are ready.

I'm ready now. Well, I'm going to simply talk.
Well, I'd like to thank you all for coming.
First, excuse me if sometimes it happens that I
misspeak the words. It is a foreign language to
me and my stock of words could all of a sudden
disappear. You'll hear me searching for a word.
Please bear with my language.

I have been so surprised to see the enthusiasm in
America about this language of ours. What I'm
going to talk about tonight might help you
understand why I am surprised because this was
from the start a very humble project, in a very
small town, in a very small country, where it
usually is quite cold, which might be one of the
reasons why we are sitting so much indoors at our
computers of course.

It started, in fact, in ... well, I might add one
or two words about myself. I am hired by her
Majesty, the Queen of Denmark, to teach
Mathematics at a small college near the German
border. When I say that I am hired by her
Majesty, the Queen, it is actually true. I have
a contract with her signature on it. Now you know
how small our country really is. Her majesty does
not go out each night and lock the door to the
kingdom, as it says in Hans Andersons fairy
tales. But, she can sign contracts, personally,
with some professors, at some colleges, in some
positions. So, she expects me to teach
Mathematics, and that's what I'm doing.

Besides that I have to give courses in Computer
Studies, specifically with applications of a
mathematical kind. And that was how it all came
about. In 1972 we got a Data General NOVA 1200
mini-computer, which some of you might know or
have heard rumors about. We were supposed to
start this new subject, Computer Science. Since
I had long time ago taken a course at the
University (only three weeks, but they didn't
know that), I was expected to take care of this
new thing, and try to teach the students how to
use it.

The language that came with it was called
Extended BASIC. One of the first things that we
experienced was that it was so crowded with bugs
that even the cleaning person began to complain
about it. It was real bad. It broke down ... it
was a so called core share system - no disk. It
was only paper tape, all of it. Paper tape and
teletypes. And it broke down regularly every four
hours, so we had to re-install the programs by
reading the paper tapes at a speed of 10
characters per seconds. It was quite a job. I
think even this one would seem fast compared to
that [referring to a Commodore Disk Drive]. This
is, as you know, the worlds fastest tape station.

Another thing that seemed to go wrong was the way
the students performed. That was to put it mildly
... miserably. The programs, even small programs
were very bad. They didn't understand much of it,
and it happened again and again that they could
do it much easier on their pocket calculators.

Also, for myself, I found it extremely difficult
to mark their works, because it was hard to read,
simply to read their programs. Twenty or thirty
line programs were hard to find out what was
going on. I had to draw diagrams very often,
simply to find out what they had did, what
methods they had applied, what approach they had
used. Of course, it might have been me that was
not good enough, but that is an uncomfortable
idea, that you're not good enough, so after
awhile, I went to a nearby university, I knew
their Computer Science section was a very good
one ... in fact I went there because we had to
set up some exam papers, some examination
problems, and I didn't know what to do about
this, because I could see disaster ahead, even if
I only gave them simple problems to solve in this
language, this so-called language, BASIC.

Here, in the afternoon, I met a man, his name was
Benedict Lofstedt. When I came in and explained
to him what this was all about, he said: "Well,
this is not your fault, this is not the students
fault, it is the language that you are using. It
is in fact, not a computer language. It is an
offense to the human mind."

I am a mathematician, I'm not a computer
scientist. And we were quite impressed by this
new technology. It cost a whole lot of money and
the local newspapers had written about it. So, I
was surprised that he asked me to go back and
read a new book by Niklaus Wirth called
Systematic Programming. In that book he presented
a language called Pascal. So we looked at it and
I could see at once that this was what
programming ought to be about.

Now, we couldn't use that Pascal directly. As it
was, it was a very good language for advanced
University students and professionals. But for my
students, it would be impossible.

There are one very good thing about BASIC ...
it's interactive. You can type in small programs,
have them run at once. You won't have to write a
program using an editor, having the text stored,
have the compiler load it, compile it ... going
through all these phases of editing, compiling,
and manipulating the operating system; that would
be far too complicated for the students I had.
So, we would have to maintain some of the user
friendly environment there is in BASIC. I think
that is the only good thing you can say about
BASIC, it's the user environment. And it is
important to understand that there is a
difference between the language itself, and the
operating environment this language is working
in. We found out that when we tried to define the
language.

Benedict Lofstedt, the computer scientist, and
myself wrote letters to each other over a period
of half a year, and tried to define another sort
of language. We took what we had of BASIC, and
added the structures of Pascal. So it is true,
that we took the best from BASIC and the best
from Pascal.

It would be very hard, I knew, to have it
implemented. One thing is to define a programming
language. We had finished the definition in fact
by the beginning of 1974 - we had the first
definition, of what later became COMAL 75. But to
implement it is quite a different story. We went
to one of the major companies and they asked me,
rather cynically by the way, unfriendly, to come
back again with $50,000; then they might consider
the whole project. Otherwise we could go ahead
with what we had. No understanding at all.

Well, I then had what you would call a really
good fortune. I got two very bright students. I
did machine coding at the NOVA, which was in fact
a 16 bit machine, and I told them how to code,
machine code. In a few weeks they became much
better at that than I was, than I ever became.
They were young people, and had learned it very
fast.

One day one of them came up, and said: "This
definition of yours that Benedict and you have
been doing. I think we can do that."

I was very happy and surprised. Well try it.

They worked on it for about half a year, and
eventually came up with the first implementation.
We had a very close co-operation. Again and again
we took discussions on how to implement the
different details. I can still remember the day
that we tried to put in long variable names. This
was a problem then. There were no long variable
names. Instead of saying number, name, we could
only say N or N1. Still I think there are many
BASICs that only allow short variable names.

But we wanted them, wanted them right away in the
first version to be at least 8 characters. We
wanted to allow 8 characters. At least we would
then be able to give variables sensible names.
Such that the programs would be much more
readable. This one of the first things we were
after.

We found out somehow, that by putting these names
in arrays, and using tokens to address them, and
thereby addressing their data fields, that we
could do it that way. This sort of discussions
were going on for half a year. We had a huge
board where we wrote down all sorts of ideas. It
was a great experience. It was quite fantastic,
to be in that process with these two extremely
clever young boys.

I had a problem with one of them. He was a heavy
drinker. Now and then he would go away, and we
knew where he was. But only at a certain time
when he had had enough could we pick him up. We
could not persuade him to come back again until
he had taken so and so much. He would then be
very drunk. And we needed him, as he was our
expert in debugging. Danish beer can be quite
heavy, quite strong. We poured coffee in him, and
put him into a cold shower. Then he would work
for 24 hours. It was just great.

The first version of COMAL ... still we hadn't
called it that, we had a Danish name for the
project ... was running on the 5th of August,
1974. It was a five line program with a REPEAT
... UNTIL. And we had all the automatic
indentation ... that was all done. We were
working on the IF .. THEN .. ELSE with
indentation. We pinned it on the board, and it
was stolen a fortnight later, so I don't know
where it is. This is quite a pity.

In February 1975, we had a version to send out.
Now, you should imagine that we would send it out
on a disk. We had no disk! No diskette, no hard
disk, no soft disk, or anything. We sent it out
on paper tapes this size.

I don't know if any of you have worked on a
teletype.

* Several people cried out Yes!

OK ... you know what it's like. Most would have
a high speed reader, typically, 500 characters
per second, which was quite respectable. But many
of them would have to use a teletype 10
characters per second... it took one hour to load
COMAL. So please don't complain over this. (God
bless the poor fellow who then switched the whole
thing off by some mistake).

Again and again we went around ... by the way,
when we implemented, we only had paper tape. We
had a fast paper tape punch and a fast paper tape
reader, but we used miles on end of paper tape.
Sometimes were walking around in paper tapes up
to our knees. My job most of the time was to wind
up ... I invented a very fine technique. I think
I could even do it today if I had a chance ...
wind it up very fast.

One thing that might happen eventually, you
probably know that ... I need to tell you that
the whole implementation cost me $300. That was
all we had, so we did it in our spare time ... we
thought it was great fun. And it was. This was an
experience of a lifetime, to be doing such a
thing. It was absolutely fantastic.

When we were in the process of reading in a paper
tape, it happened eventually that one stepped on
it. And we could start all over again, because
the paper tape broke. This was one of the deadly
sins, I don't know how many are allowed.

We were fortunate, because most of the
institutions that were of our type, (high
schools, colleges, vocations schools, etc) had in
fact Data General NOVAs, because one local Danish
company was the distributor of that machine. We
didn't have problems with such obscure machines
such as Hewlett Packard or whatever that is. We
had the NOVA all over. So they took it and
started to use it. We were quite successful. I
think it took half a year and they had stopped
using the regular BASIC and turned over to COMAL.

Now, as it appeared, the major reason for doing
that was the long variable names, believe it or
not. Especially the students. The teachers were
more hesitating, but the students took it very
fast. They could now write NUMBER=5 or
NAME$="Peter Sorenson".

And then the IF .. THEN .. ELSE. That was one
that really was bought at once. I don't know if
you are aware that doing branching, especially
nested branching, using GOTO is in fact very very
difficult. Therefore we had the nested IF ...
ELSE ... ENDIF only up to a level of four at that
time, but that appeared to be enough.

And then of course, the third facility, was the
named procedures. We didn't have parameters at
that time. You might call it simply named
subroutines. Instead of saying GOSUB 5000, we
could say EXECUTE PRINTOUT. And then you would
have PROC PRINTOUT ... ENDPROC. That was all
there at that time.

This was all much much easier. And it was faster.
This was another thing we noted and which helped
us. I don't know if you are aware that when BASIC
tries to find line number 5000 starting at line
number 10, it will go over all the line number to
try to find it. In the first COMAL we did, we put
up links, such that if there was an EXECUTE (that
was the keyword we used) EXECUTE PRINTOUT in say
line 50, and PROC PRINTOUT would appear in line
500, there was a pointer telling line 50 to jump
at once (in machine code) to line 500. That of
course was much much faster than having to go
sequentially through all the line numbers. It was
also one of the advantages that made COMAL
popular very fast.

This went on, and I gave it the name ... in fact,
I can tell you how that happened ... the name
COMAL. We had been discussing for a week what we
should call it. We couldn't use the Danish name,
we were well aware that we should have some fancy
English looking name, otherwise no one would have
it. In fact, I was on my bicycle from the college
back home when it appeared to me, that we had
been fooling around with the language ALGOL,
which you probably know, ALGOrithmic Language,
and all of a sudden it appeared that this should
be COMMON Algorithmic Language. Shortened is
COMAL. That was in fact how the name was made.
That's all. It is a really country name.

We continued to improve it a little ... as I said
before, we then sent it out. I was not doing much
more machine coding, because the students asked
me to write the documentation. That's of course
a very important thing. I also wrote a text book
which later became Beginning COMAL. That took 3
or 4 years before it developed into the Danish
edition of that book.

This went on until around 1978, when the first
microcomputers appeared. We got a problem. The
problem we had to face would be the following.
Imagine the Danish schools had started to buy PET
or APPLE. That meant that we were sent back to
BASIC again. This was to me very frightening. It
was so frightening in fact that if I had been
forced to teach BASIC again, I would not have
taught computers at all. That's true! I was
completely determined, if they were to force me
to teach BASIC again I would have said: "OK, you
hired me to teach Mathematics, Algebra,
Statistics, and Number Theory, and I'm good at
that, so that's what I'm going to do. You can
take your computers somewhere else." I was
absolutely determined to do that. But of course,
it would be better if we could come up with a
microcomputer with COMAL.

So in 1978 I defined a new version, where we put
in the parameter mechanisms and the sort of
things you know now, most of it at least. And we
found another group in a college somewhere else
that were willing to try to improve, enhance,
expand COMAL into what we call COMAL 80 because
it was not until 1980 when it was really
finished, and it was running on a Z80 based
computer, Zilog 80. So we had two reasons to call
COMAL 80.

At the same time another Danish company designed
and built a computer for schools. It was called
the Comet. We are not always that modest in
Denmark. It appeared to be a very good computer
by the way. It is still used especially at
technical schools because it has one of those
busses where you can add in extra electronics all
the time to the buss. It is still very popular.
That was the first computer to run a COMAL that
you would recognize. If you saw programs written
for that computer, you wouldn't be surprised at
all. This would look very much like what you see
today. What had to come would then be to make
better environment, better editors, to put in
extra facilities, such as graphics, sprites, and
sound.

But at that time, 1980-1981, I stopped doing much
about it. It was then taken over by different
groups, one of which is UniComal which is one of
the most talented. Maybe I should add now about
the way UniComal started. One of them Mogens
Kjaer, wrote a letter to me, where he criticized
my definition of COMAL 80. When I saw this
letter, I knew that once that he would be one of
the boys that could do something about it. This
was simply one of the most qualified critics that
I ever had. It was very good. I adjusted my COMAL
according to what he advised me to do. And then
I asked him if he might do ... he had intended to
do an implementation for a computer that wouldn't
have any chance of going anywhere. I advised him
not to do that, but instead to go for the
Commodore PET. It was obviously a very good
computer.

At that time, you should know, by the way, a
company could not sell a computer to a Danish
school if it did not have COMAL. Hewlett Packard
never sold their computer since it did not have
COMAL. We had given it such a strong position.
The teachers wouldn't have it. The teachers
refused to use BASIC. That was a very strong
situation. We hold in fact a very strong position
without telling anybody. Simply by giving them a
language that was apparently better.

Mogens Kjaer made this version for Commodore, and
this is the version now known as 0.14. After he
did that, the UniComal group was set up and they
then went on. From then on I have not had much to
do with it. These young fellows are doing a
terrific good job. They added the packages, which
is probably the most important extension to COMAL
that was made after the latest definition. This
was the package concept that you could have
packages in machine code and call them.

From then on you know more or less what the story
is all about. Still a new version now is being
designed for the Z80 CP/M based computers for
there is an English computer called the Amstrad,
which is quite popular in Europe. They expect to
sell around 600,000 of them. We have a version of
COMAL for that now [this is the CP/M COMAL]. It
is very much like the UniComal version. There are
a few improvements, but not much.

What I'm interested in now ... I don't think we
should expand COMAL much more. I feel perfectly
well with the version for the Commodore 64. I
have 17 of them for my students. We will probably
have one or two PC's only because an institution
like ours must have PC's. I probably can persuade
someone to use them if we get them.

What I'm interested in now is to see a better
user environment. It has improved very much with
FIND and CHANGE. I would like to see full screen
editors, without line numbers. There is a COMAL,
Mytech [Alder] COMAL, with such a very good
editor, but I'm sorry to say that the COMAL is
not as good as it ought to be, but that may come.
There is now an international standardization
committee. We have a meeting in Scotland in
September. We are going to discuss, among other
things, whether we should get rid of line
numbers, or at least make them redundant. You can
get them if you like, but you can avoid them. You
are aware I expect, that line numbers are not
needed after we have left the teletype. You can
with a screen editor, you can move the cursor
wherever you like, and insert a line by
manipulating some control code. Whereas with the
teletype, line numbers were quite ingenious. You
couldn't get back to the place between line 10
and 20, but you could print line 15 and have it
inserted between line 10 and 20. This was in fact
very well made considering that it was made for
teletypes. But they are, I think, gone long ago.

This is more or less how it was. When we started
it, we ... well if anyone had told me in 1974
that 10 years later I would be standing in Los
Angeles, I probably wouldn't have believed him.
Because we only wanted, as simple as that, was a
tool that our students could use. Nothing
pretentious or great project or anything. It
appeared that our students could use it, it has
developed over the years. It is therefore, as you
may guess, very impressive, and makes me ... Well
I don't know exactly how happy I really am, maybe
I'll find out when I come back, maybe, that you
really like it and are so enthusiastic about it.
I think it is a good tool. It appears to be so.
And I hope that many more people can get to know
about it.

Well, you are welcome to ask questions and ...
that's a lot of detail, but I may add that I used
a slogan ... I hope you don't think I'm nasty ...
I used a slogan that in the 70's I said that: "It
might very well be that the good God has invented
the computer. But there is no doubt that his
black majesty then came up with BASIC shortly
thereafter [laughter] ... in order to confuse
peoples minds, because that's the way he prefers
to operate." [applause].

* I would like to begin by taking a poll. Of all
the COMAL programmer here, I would like to hear
yays or nos from those who would like to see line
numbers implemented for COMAL 2.0. All who would
like to have line numbers ...

* [interruption] What about how it is to be
implemented? If you remove line numbers, how do
you differentiate being in command mode and being
in program mode.

That's right. Now, I think I said that it should
be made redundant. But you would have to have two
modes, an editing mode, and a runtime mode. In
fact you can see it on the Mytech [Alder] COMAL,
how it operates. There's no problems. I can
demonstrate it after this meeting for whoever
likes to see it. This is very easy to implement
and use. But still you can use line numbers.

* What would be the object of getting rid of the
line numbers? What is the point? What is the big
controversy of getting rid of them?

I would hate to admit that it is some academic
idea. But I think it is. The idea is ... when I
say full screen editor, I do not mean only the
text you can see on the screen right now, but you
should be able to scroll up and down. You can now
in COMAL have the listing coming up, but you
cannot take the cursor up and have the listing
coming down again [note: this is now possible in
IBM PC COMAL, AmigaCOMAL, and C128 cartridge
COMAL]. So you should be able to look at the
screen as simply a window, through which you pull
a film either forward or backward.

But that is not the most important thing. I think
I am best at giving examples because I am not a
computer scientist, so this is not ... I am not
one of those desktop academics that think these
things out. I look at the students and try to
find out what they need. And then we do it. COMAL
is very pragmatic. We never took in new
facilities and new structures until we felt that
the students needed them. If they made a lot of
mistakes, or if they had to do more work than
necessary with details, then maybe an extra
facility would make it much easier for them. That
is in fact how the parameters were introduced. We
discovered of course, that parameters would make
it much easier for them, so we introduced them.
But not until then.

Now, imagine that you have a list on the screen.
Say from 10 to 240. You want to insert a line
between 30 and 40. No, I think I'll take a better
one. You'll have a list from 610 to 700, or, my
mental calculation is a little ... that is why I
am a mathematician, I hate numbers. As a child
when I found out that you didn't have to use
numbers, that you could use letters instead, that
was it!

OK, you have here some numbers from six hundred
something to six hundred something else. You want
to insert a line. You know then, if you have to
use the line number. Very easy. Instead of saying
615, you write 15 for example. You introduce a
wrong number. Later you cannot find that darn
line. Finally you see it. What is doing there. It
is between 10 and 20. It should be between 610
and 620.  What you should be able to do, with or
without line numbers, is to move your cursor up
to 620, press CTRL I - the text opens up, and you
insert the line. So you get what you see. That's
the whole idea. Also, if you put the cursor at
601, press CTRL D, and zip, the line disappears.
This is what I am after. This is what I am trying
to get the standards committee to accept at the
next meeting. I know I'll have a hard time doing
it because I know it is very difficult.

That's another thing I might have told you. I got
into an enormous lot of fights and arguments when
I came out with the language. In fact, don't
expect that people are happy and grateful, for
your coming up with this. A lot of people would
say: "Why what's this about, we are perfectly OK
with what we have."

There were articles in papers written against me,
as though I was some new unknown form of
Communist in disguise. It took me some years to
understand why people can get mad at you when you
come up with new ideas.

A programming language is not just a tool. It is
something different and something more. In fact,
what I was saying to some of these people ... or
how I was interpreted was: "You are telling us
that we are using bad language." And that is
certainly not a nice thing to tell people.
Really! It is! Now I am much more careful not to
be too tough. You think the saying comparing
BASIC and the Devil was bad enough. Wow. You
should have heard the arguments we used in
Danish. They are absolutely untranslatable. They
never should be anyway.

I hope I answered the question about the full
screen.

* What you are really after is a full screen
editor. The line numbers are neither here nor
there.

Well OK. In a 40 column screen it would give you
5 characters more. And sometimes it may prevent
the wrap around. But that is minor. And whether
the line numbers are going to be there or not is
also of a minor concern. The important thing is
that you have a full screen editor. You get what
you see.

* Referring to that problem, not only not being
able to find the line that you were trying to
insert when you mistyped it. About three days
ago, I had one of those. I could not find it.
Nothing at all worked. What I did was I double
struck one of the keys and it typed in place of
a line I had already written and everything went
to pot.

Yes, that's right. If you could open up the text,
you would know exactly what you were doing. You
would see which line was pushed in order to make
room for the next.

Well, I've seen a lot of this with my students.
Coming up with a program and saying, well I had
a program here 15 minutes ago, but ...

I remember, by the way, one experience we had
with BASIC programs, and this is a true story
again. Long John was the students name. He was
almost 2 meters. He was a favorite with
basketball. He was also a good programmer, but he
came up one day with this, not a very big
program. He asked me, in Danish of course, in a
fair translation, he asked me the following
question: "Borge, could you explain to me why
this program works?"

Well, John. We sat down and we found out that
three lines were never executed. He jumped around
them. We sorted things out ... that was in the
old bad BASIC days. He was only one out of a
series of similar experiences. Now, if I should
put a ... I could hardly use the good English
work BASIC any more, but a basic philosophy, it
would be the following: "Computers should be as
easy to use for human beings as possible."

We should be controlling computers, not the other
way around. This was what was behind us all the
time when we made this [COMAL]. Benedict Lofsted,
the first one ... he was not in the second
project because he went to Greenland, and that's
a long ways from Denmark to Greenland. So, I had
to the second project all alone. But it was not
that difficult, because it appeared that the
first definition was so well designed that we
could enhance it, extend it, without violating
the original idea.

This is one of the other things, I hope you
agree, that it still appears that COMAL can take
extensions without the original concept being
ruined. Maybe that is why you like the language
in fact ... is the regularity in it ... or as a
computer scientist would say, the orthagonality
of it. I tried to learn that word.

* One problem I have as my program gets growing
to some length, that you start forgetting
procedure names and things like that. Is it
possible to have a window or something which
could list the procedures that you have going at
this point?

Yes. What you are asking is a list of the symbol
table, of the variable names and the procedure
names, probably with some indication of where
they were to be found. We have that in the new
CP/M COMAL. We have a LISTSYM or LISTPROC, I
forget what we called it. Note that this has
nothing to do with the language. This is the
environment. And that is what is what we are
after now. To make it even more friendly.

* Is there a possibility of having certain
procedures built in? For example, rounding up or
that sort of situation?

Yes, it should be. That is what they are doing
with the machine codes, the packages. In fact,
you can build in anything. You can make yourself
a commercial package.

* Also, like COMAL 2.0 has built in graphics, if
you want to draw a circle, you just use the
keyword circle and its parameters. In 0.14 you
have to have a procedure.

In the Mytech [Alder] version ... I hope it will
eventually be a good version. You can even write
the code [for packages] in the language C.
[AmigaCOMAL also allows this] And C is much
easier to use than machine code, in case you
don't know it. And then use them as COMAL
packages. That is to say, that all the procedures
you write in machine code can be invoked or
called as if they were normal user defined
procedures.

* I wanted to tell you a couple days ago, but I
never got around to it. There is an environment,
called artificial intelligence, developed at the
University of Suffax, called PopWalk. It works in
a screen oriented environment, providing you
access to LISP, PROLOG, and POP11... that's the
three languages. What I thought was extremely
powerful, is that you could link to it
subroutines, procedures, whatever you want to
call them, in a host of other languages,
including Pascal, Fortran, C. That strikes me as
being extremely powerful. Each of these languages
have some particular advantage, for some
particular applications. Maybe it's the language
you happen to know, that's already an advantage.
Otherwise the language may be well suited to
doing a particular task. So, I really try to
encourage the idea that COMAL should be able to
access routines compiled in a wide variety of
Languages.

I don't know whether that could be done, or how
feasible it is, but it would be very practical to
have it [in 1989 AmigaCOMAL and IBM PC COMAL 3.0
now allow this]. Of course you should be aware,
that a language cannot be totally general. There
are applications where even COMAL wouldn't cope
with. We had a discussion with Seymour Papert,
the man who made LOGO. He put out his position,
quite clearly that, ... by the way, he took a
look at COMAL, and he liked it! ... and he
expressed his opinion that even children should
try to learn more than one programming language.
It will be more like programming concepts that
they will be learning about. And it appears also
that children are learning these things extremely
fast. The robots, which I've brought one or two
samples with me, you can look at them later if
you like, were programmed by 11 year old
children. We had no problems. The teacher had
problems. The children had not.

* Well, it seems, from what we've learned today
of programming languages, that there are really
three classes of languages. Each of them fulfill
a role that the other ones cannot.

Yes. My guess is that in the future there will
only be two. I think that the LISP languages will
be taken over by the Logic, the PROLOG type. I
think so. The three classes you think of are
probably the algorithmic languages, the LISP kind
of languages, and the logic, PROLOG, kind of
languages. COMAL is of course, an algorithmic
language, it's in the name. LISP languages would
be LISP or LOGO, LOGO is a LISP language. The
third one would be PROLOG, a language that is
mentioned very often, especially when you speak
about 5th generation computers, and the
competition between the United States and Japan.
Who will be there first with this super thing.
But I think that eventually you will see that
there will be two. The algorithmic languages,
they will never disappear. In robotics, they will
still perform. Just imagine ADA. ADA is an
algorithmic language, which by the way, COMAL
would be a very good kindergarten ADA. It could
be an introduction to ADA. I really think that.
Pascal is an algorithmic language, Algol, etc.
Whereas PROLOG,... and I think it is with PROLOG
... PROLOG has much the same relation to LISP, as
Pascal would have to FORTRAN.

Well, the GOTO... Of course, in COMAL there are
of course lots of GOTO, but they are hidden. The
interpreter is doing them for you. IF, ELSE,
ENDIF ... from IF you can jump to ELSE if the
condition is not met. This is of course a GOTO,
but it is hidden. Because it is being expressed
in a way closer to the human language, or the
human way of thinking. And the same way I guess
that PROLOG would be able to handle lists in a
way that would do the same to the list structures
that the algorithmic structured languages have
done to GOTO. That's my impression. But, OK. I'm
sure of one thing. The algorithmic languages are
there to stay.

* Fortran even.

Yes, I think so. But, OK. COMAL is first of all
for people who are not professional. I think I
wrote in the handbook, to make it possible for
people to program computers even if they were not
programming people, or something like that. And
it appears that it can also be used for some
semi-professional, maybe even ...

* [interrupts] Paradoxically, I think that the
professionals are the fastest to appreciate it.

So much the better.

* Of course, the trick is to get their attention.
Once they see it, a little exposure to it, they
immediately see that it is just the thing.

This is probably the most important point you are
indicating there. I heard one tonight who tried
to introduce COMAL to some teachers somewhere.
They have to see it to believe it. They must have
a chance to see what COMAL is all about, to see
a few examples, a few real programs, before they
use it. You can give long talks, even very
delicately formulated talks, about it. But there
is nothing like a short example. You will see
that, if you want to tell people about COMAL, you
will have to show them a program in COMAL, and
let them see the syntax, the error messages they
get immediately, and the indentation that shows
them... Let them look at an IF .. THEN .. ELSE,
and REPEAT .. UNTIL, simple things, procedures
can come later. As soon as they have seen that,
they will know what it is about.

Now, I am running out of words. I can feel it.
But don't hesitate to ask questions if you like.

* Being one who has almost next to no experience
in COMAL programming, I was interested if you
could elaborate on those three levels of
programming language that you just mentioned.

I could, but I am afraid it would be too
technical. Do you know about LOGO and PROLOG. I
would gladly do it afterwards. It is quite
technical. The only thing I can say is, COMAL is
a very typical algorithmic language. And if you
want to know about a LISP language, the easiest
way would be to look into LOGO, not at the turtle
graphics! Because that has really nothing to do
with LOGO. You can do turtle graphics in any
language. You see it in USD Pascal for one. That
was a real sensation. That he [Kenneth Bowles]
used turtle graphics to train university
students. And probably what made USD Pascal so
popular.

You'll have to look into the data structures.
LOGO is not using arrays like COMAL is. DIM a(10)
or whatever it would be. LOGO is using so called
lists, which are more or less what mathematicians
would call sets. PROLOG is totally different. It
is the language that uses relations. It is more
or less abstract algebra.

* I have a simple question. Could you relate all
the computers that you know that run COMAL now.
We'll give you ten minutes.

Oh no. That would be boring.

* No. I understand that there are some foreign
computers. I would just like to know.

Well, in Denmark we have the old Z80 based
microcomputers, which would look very much like
Cromemco or IMSAI or whatever. In fact, Comet
would look very much like the IMSAI, although the
IMSAI used 8080 as far as I know. Then there are
new ones coming up now, based on the 8086.
Piccolo, and another one called Partner. More or
less PC clones. And the Swedish Compis, which is
the one Mytech [Alder] COMAL could be used on,
also a 80186 based computer. Then the English
Amstrad is one you may not know about. Apple has
a COMAL in fact, Metanic COMAL. You have never
seen that. It is not very good, I would not
recommend it. A new one is being implemented
right now. But only if the Apple has a Z80 card
(CP/M). Then there will be PC COMAL of course.
And the VAX computer. And Wicat. I don't know
much about that, only that there is a COMAL for
it. And Mytech [Alder] claim that they have one
for the Mac. They call it MacCOMAL. Whether that
is true or not, I don't know, I never saw it. [it
is true ... we have seen the beta test of it
working]. And I hate to speak about things I have
not seen. Because in this branch, rumors ...

* Have you considered introducing a help facility
in the COMAL environment.

Yes, Mytech [Alder] has a HELP. And the new Z80
version has a HELP. In fact the first test
versions were running right before I left
Denmark. This will be another thing that will be
taken up by the Standardization Group. It is more
or less the environment again. And it would be
very useful I agree. I hope that COMAL will go on
developing. But you should expect it to be more
in the environment. The language as such I think
is probably more or less finished. The definition
is closed. There will not be added many more
facilities. There are a few things, such as
functions. You know that in 0.14 a function can
return integers or reals, but not strings.
Whereas in 2.0 functions can return strings. In
Mytech [Alder] a function can return vectors,
which might be an improvement in the language
itself. Otherwise it would be to make it more and
more user friendly.

* What is the Mytech [Alder] version based on?

It is written in C. This means it can run on any
computer that has a C compiler. Probably it would
be very easy to do, if they haven't got one
already, a 68000 version. We are trying to
persuade them to put it on the Amiga, whenever
that appears.

* I've seen advertisements for PROMAL. Is that
related to COMAL at all?

Not at all. I saw a German review of it. In
Germany COMAL is very well know and is used
extensively in many schools. The reviewer started
to point out explicitly that PROMAL has nothing
to do with COMAL.

* Coming back to a point raised earlier. How
about the possibility of compiling functions and
procedures written in COMAL; compiling them
directly into machine code. Once it works, it
works forever, and then it runs a lot faster.

That would of course be a thing that probably
anyone would like to have, a COMAL compiler. You
would have something like the perfect system
then. The interpreter to develope interactively
the program, to test it. As soon as the program
runs as it should, you would take the program and
compile it. It should be very easy to do, and I
hope that someone will do it.

* Let's let Borge get some rest.