Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Stalin / New tutorial

8 views
Skip to first unread message

phi5...@yahoo.ca

unread,
Mar 19, 2008, 11:58:39 PM3/19/08
to
From the number of downloads, and emails received by members of my
group, I noticed that stalin is much more popular than bigloo. I
cannot understand the reason for this strong preference for stalin,
since bigloo is also a fine implementation of Scheme. Anyway, I posted
two more chapters of the stalin tutorial. I give many examples of
graphic programming in the new tutorial. I believe that my examples
are shorter and easier to understand than the sole example of GUI
programming that comes with the distribution. My examples run both on
Windows, and Linux. BTW, the new tutorial is called beria.pdf. With
each new posting, the old tutorials become obsolete.

klohm...@yahoo.de

unread,
Mar 20, 2008, 2:01:52 PM3/20/08
to

It would be interesting to know how many people are actually using
Stalin or Bigloo on a daily basis. I have used Bigloo for nearly over
5 years on a daily basis.

I am not convinced that download statistics tells anything. Sometimes
people are playing around with programming languages.

But anyway the more there is activity within the Scheme community the
better.

What I am much more concerned with is the fact that some Scheme
implementations, eventhough alive for over many years, are driven by a
single person so to speak. This is true for Bigloo and Stalin. How to
convince people to write a huge amount of code in such implementations
especially with regard that fully fledged Bigloo programs (which
really try to employ a lot of Bigloo extensions) are not as easy to
port to other distributions of the Scheme language than lets say a
standard Common Lisp program.

The thing I like most about Bigloo: it spots a lot of errors at
compile time. I am not in the static type bussines but I must honestly
say I was a bit disappointed when I tried my bet on Chicken and was
really surprised how hard it would have been to program in Chicken
where I will get error messages just at runtime. Chicken is surely a
great effort and I think the community is even greater than the Bigloo
its one but it was the first time that I could a little bit understand
when people start to hating Scheme for just the pure fact to wait so
long up to hitting ./a.out for runtime to get the first error
messages.

phi5...@yahoo.ca

unread,
Mar 21, 2008, 1:43:57 AM3/21/08
to

I couldn't agree more with you. I mean, I agree with everything you
say about Bigloo, and Stalin. I personally prefer Bigloo to any other
Scheme. As for using Bigloo in a daily base, I believe that there are
many very productive programmers doing so. For instance, the people
who maintain the roadsend php compile (http://www.roadsend.com/home/
index.php), Dr. Romildo Malaquias, who created the integrated
computer algebra system (http://www.decom.ufop.br/prof/romildo/icas/),
Will Farr, Jorlano, to name just a few people. By the way, the new
numerical tower of Bigloo comes from a library that Dr. Malaquias
created for his computer algebra system.

The situation of myself and friends is the following: We started
college not long ago. Most of our professors prefer Java/Python/Matlab/
Labview; one of the most influencial and productive professor uses
Clean, but he is an exception. However, a few students, like myself,
think that Scheme is more concise, fast, and productive than Java,
even for doing assignments and homeworks for teachers that adopt Java/
Matlab; this specially true in course like Data Structures, Compiler
Construction, and Artificial Intelligence. Of course we pay a heavy
penalty for choosing Scheme over Java/Matlab. Here are the rules that
we must abide to:

1 --- Beautiful graphic user interface. As you can see, we are
creating a GUI to use with Scheme. Since Matlab can be used to plot
things, we need similar resource for Scheme.

2 --- Small static exec files. Although neither Java nor Matlab
generate small exec files, and the TAs accept programs in those
languages, they do not let us install dll/so libraries in his/her
machines; one rule for us, another one for people who program in
Matlab. Therefore, we need to create static exec files small enough to
email. In general, the university email server accepts up to 2M bytes
of attached exec files. Both Bigloo, and Stalin satisfy this
requirement; DrScheme does not, as far as I know.

3 --- Our programs need be faster than Java. Since we are not using
Java, the TA of NA requires programs that are faster than the Java
equivalent. Interesting, but TAs accept programs in Python, and
Matlab, that are very slow; again, one rule for us, another one for
people who program in Matlab; they claim that Python and Matlab are
mainstream languages, cherished by the industry, while Scheme is a toy
language; we have the burden of proving that Scheme is not a toy
language. Don't worry: Java is very slow; there is no difficulty in
creating programs in Bigloo/Stalin that outperform any Java code from
other students. If you take a look at our Bigloo tutorial, you will
see a few tricks that we use to make sure that our programs outperform
Java (see Chapter 8/vectors); the main difficulty is array processing;
we have used define-syntax to create loops that maps 2D arrays on
homogeneous vectors, as described by Farr.

I have an interesting story to tell you. My cousin is fond of Prolog.
He had an assignment, where he was supposed to write a piece of
software for doing schedule analysis. Of course, he wrote the
assignment in Prolog (Visual Prolog, I believe). I cannot tell you
why, but his program outperformed the code of the other students, who
worked with C. Maybe he used better algorithms, maybe Prolog is good
for this kind of thing. I do not know the reasons. The fact is that
his program was fast. The TA kept rolling his source code on her
terminal, recompiling it, and rolling the source on the terminal (2000
lines) again, with a look of disbelief on her face. Finally she
accused him of plagiarizing, but since she could not prove it, she
gave him the minimal mark he needed to pass the examination.

ren...@gmail.com

unread,
Mar 21, 2008, 3:43:54 AM3/21/08
to
>Anyway, I posted
> two more chapters of the stalin tutorial

What's the URL? Thanks!

Renick

ren...@gmail.com

unread,
Mar 21, 2008, 3:46:50 AM3/21/08
to

Sorry to reply to myself... the URL is in this thread from a few days
ago:

http://code.google.com/p/stalingui/

ren...@gmail.com

unread,
Mar 21, 2008, 3:47:01 AM3/21/08
to
On Mar 21, 4:43 pm, "ren...@gmail.com" <ren...@gmail.com> wrote:

Sorry to reply to myself... the URL is in this thread from a few days
ago:

http://code.google.com/p/stalingui/

ren...@gmail.com

unread,
Mar 21, 2008, 3:55:52 AM3/21/08
to
On Mar 21, 4:43 pm, "ren...@gmail.com" <ren...@gmail.com> wrote:

Sorry to reply to myself... the URL is in this thread from a few days
ago:

http://code.google.com/p/stalingui/

klohm...@yahoo.de

unread,
Mar 21, 2008, 6:24:14 AM3/21/08
to

I have also been using Clean in the past. Clean was my only experience
in the field of 'real functional' programming languages.


phi5...@yahoo.ca

unread,
Mar 21, 2008, 6:25:01 PM3/21/08
to

If you want to take a look at our Clean site, the address is

code.google.com/p/CAW

You will find basically the same material that we posted in Bigloo.
However, Bigloo seems to be more appropriate for GUI programming.

Rob Warnock

unread,
Mar 21, 2008, 10:24:42 PM3/21/08
to
<phi5...@yahoo.ca> wrote:
+---------------

| > I have also been using Clean in the past. Clean was my only experience
| > in the field of 'real functional' programming languages.
|
| If you want to take a look at our Clean site, the address is
| code.google.com/p/CAW
| You will find basically the same material that we posted in Bigloo.
| However, Bigloo seems to be more appropriate for GUI programming.
+---------------

Uhh, URL paths *are* case-sensitive! Try this instead:

http://code.google.com/p/caw/


-Rob

-----
Rob Warnock <rp...@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607

0 new messages