www.digitalmars.com         C & C++   DMDScript  

D - IMP - another Martian D effort?

reply Mark Evans <Mark_member pathlink.com> writes:
I just had to post this because of IMP and because of their Mars theme. ;-)
Is IMP the same type of language as D?

http://www.redshift-software.com/
http://implanguage.com/docs/Intro.html

"IMP is low level, imperative programming language intended for approximately
the same kind of tasks that C was originally designed for i.e. system
programming, tools, and moderate sized applications. IMP's  design goals are
pretty much as follows: 

A consistent syntax. 
Simplicity 
Small Language Size 
Sufficent power to build an Operating System 
Generality 
Features deliberately unlike other languages"

Mark
Oct 01 2002
parent reply "Walter" <walter digitalmars.com> writes:
Imp has similar goals, but the result (if you look at the syntax) is very
different. It also appears to have been abandoned.

"Mark Evans" <Mark_member pathlink.com> wrote in message
news:and7r7$2abq$1 digitaldaemon.com...

 I just had to post this because of IMP and because of their Mars theme.


;-)

 Is IMP the same type of language as D?

 http://www.redshift-software.com/
 http://implanguage.com/docs/Intro.html

 "IMP is low level, imperative programming language intended for


approximately

 the same kind of tasks that C was originally designed for i.e. system
 programming, tools, and moderate sized applications. IMP's  design goals


are

 pretty much as follows:

 A consistent syntax.
 Simplicity
 Small Language Size
 Sufficent power to build an Operating System
 Generality
 Features deliberately unlike other languages"

 Mark
Oct 01 2002
parent reply Mark Evans <Mark_member pathlink.com> writes:
I got that impression too.  I told them about D.  They have some good ideas.
More importantly, they dig Mars.

Mark

In article <andle5$2o1r$1 digitaldaemon.com>, Walter says...

Imp has similar goals, but the result (if you look at the syntax) is very
different. It also appears to have been abandoned.
Oct 01 2002
parent reply "Walter" <walter digitalmars.com> writes:
"Mark Evans" <Mark_member pathlink.com> wrote in message
news:andm1f$2oj7$1 digitaldaemon.com...

 More importantly, they dig Mars.


Lot of that going around <g>.
Oct 01 2002
parent reply Mark Evans <Mark_member pathlink.com> writes:
One thing worth reading about IMP (dead as it may be) is their commentary on
semantics and syntax.  It is spot-on.

http://implanguage.com/docs/Intro.html

Mark

"A consistent syntax has similar semantics for similar structures and different
semantics for different structures. Unfortunately, most programming languages
violate this simple concept in a number of ways. C, for example, has constructs
that have different semantics for similar syntax and multiple syntactic forms
with the same semantics. The static keyword is an example of the first case,
since it can restrict the scope of a variable or function, or it can specify
that a variable's storage is not allocated on the stack (and therefore survives
from one function call to another). On the flip side C has not one, but three
different ways to dereference a pointer. (using *,[], and ->) "
Oct 02 2002
parent reply Rene Rivera <grafik.list redshift-software.com> writes:
In article <ang37s$29dk$1 digitaldaemon.com>,
 Mark Evans <Mark_member pathlink.com> wrote:

Sorry for my delay in response... had to take time to read about D before
posting anything reasonable.


 One thing worth reading about IMP (dead as it may be) is their commentary on
 semantics and syntax.  It is spot-on.
 
 http://implanguage.com/docs/Intro.html
 
 Mark
 
 "A consistent syntax has similar semantics for similar structures and 
 different
 semantics for different structures. Unfortunately, most programming languages
 violate this simple concept in a number of ways. C, for example, has 
 constructs
 that have different semantics for similar syntax and multiple syntactic forms
 with the same semantics. The static keyword is an example of the first case,
 since it can restrict the scope of a variable or function, or it can specify
 that a variable's storage is not allocated on the stack (and therefore 
 survives
 from one function call to another). On the flip side C has not one, but three
 different ways to dereference a pointer. (using *,[], and ->) "



Wow... thanks for the quote. :-)

Guess I'll say some comments on the IMP language... as far as what's not on 
the web pages...

No the project isn't dead. It's in somewhat of a slumber because of time
constraints. It's just me and John working on the project. But currently
we are working on the Thot tool project. Just trying to keep bread and beer on 
the table ;-)

We actually actively talk about the language from time to time but most
of that doesn't make it to the web site, again for lack of time. We also happen
to be discussing an object oriented counterpart to Imp.

The Mars theme I think is somewhat of a fabrication on your part ;-] Not that 
we don't like Mars, in fact we love it, but it has nothing to do with the 
language. The Mars theme comes from the "Mars Chronicles" game we are also 
working on (no public website for that yet).

As for the features and objectives of Imp as compared to D. I'd say that they 
are somewhat different, reagarless of actuall syntax. I think the most obvious 
difference is that D is trying to "fix" C/C++, but Imp is trying to replace 
what C was intended to do. We are specifically avoiding non-imperative 
language features. We want a language to use for writting things like device 
drivers, OS kernels, vector algorithms, etc. Things that C, as the current 
imperative systems language, has failed to provide.

I find it interesting you picked the above quote... It is probably the biggest 
gripe we have with C/C++ and many other languages. And is something that we 
think all new languages should strive to accomplish.

To highlight the quote and to give you a direct idea of how that applies in 
Imp I thought I'd translate the first sample program you have on the D 
documentation, the prime number sieve:

----


SieveProgram:
{
    StdIO: import;
    Types: import;

    main: export routine (argc: integer; argv: &&char)(result:integer)
    {
        count: integer <- 0;
        flags: [8191]boolean;

        print(stdio.output,"10 iterations\n");
        loop iter : integer <- 1
        {
            count <- 0;
            flags <- true;
            loop i : integer <- 0
            {
                if on ([i]flags)
                {
                    prime: integer <- i+i+3;
                    loop k: integer <- i+prime
                    {
                        on (k < size(flags));
                        [k]flags <- false;
                        k + <- prime;
                    }
                    count + <- 1;
                }
                i + <- 1
                on (i < size(flags));
            }
            iter + <- 1;
            on (iter <= 10);
        }
        print(stdio.output,"\n");
        print(stdio.output,count);
        print(stdio.output," primes");
        result <- 0;
    }
}

SieveProgram: import;
----

This is as close to a transliteration as I can manage, some slight scoping 
differences, to make it easier for you to map to and from.

But the big difference is that there is only one kind of looping construct. 
And it can accomplish not only what "for", "while", and "do-while", but 
considerably more with simpler semantics than any of them.

OK, that's enough rambling on my part...

Enjoy.

PS. Mark, didn't I see your posts in the Python C++/Boost.Python list? Just 
curious if my posts there are how you found out about Imp.

-- 
-- grafik
-- Don't Assume Anything
Oct 02 2002
parent Mark Evans <Mark_member pathlink.com> writes:
Rene,

That's a pretty good assessment of D.  I wish it had more of IMP's goal,
wholesale replacement of the bad stuff in C/C++.

I posted your nice summary of C's design flaws for that reason.  It connects
directly with recent discussions on this list.  C began life as a kind of Unix
geek hack that just grew out of its britches. We have been suffering ever since.

While I argue that D should adopt high-level constructs from STL and script
languages, I still want the low-level bit twiddling of a systems language.
Recently we had some postings about D and embedded systems.

C syntax needs to be cleaned up and am glad to see someone doing that.
Personally I don't mind living with legacy C syntax if the inconsistencies are
removed and the semantics repaired.  D could do that, but there is a strong
desire to migrate C/C++ folks seamlessly, which raises the resistance level to
better language constructs.

One thing both IMP and D share in common is a desire for maximum performance and
there have also been some discussions about that lately too.

The D folks are also talking about iterators now which connects directly with
the looping issue, a rather novel thing in IMP.

The web interface to the D newsgroups permits sorting by date, author, and other
fields.  If you use my name as the key then you can read through various threads
in which I have participated. That should give you some idea of my thinking.

I don't recall exactly how I discovered IMP.  Please put some of these
philosophy and mission statements on your web site.

Yes I have used Boost Python Library since it began and keep up with the
C++-sig.

By all means keep rambling to us. I hope that you will contribute ideas to D
even if you keep working on IMP. I think your ideas are sound and long overdue.

Mark
Oct 03 2002