[Oz] Oz vs. Squeak

[Mikael Kindborg]

At 16:02 2001-01-30 +0100, Denys Duchier wrote:

> > What I am asking is if there is any interest in the Oz-community for
> > pursuing programming environment and user-interface issues?
>
>Yes, there certainly is, at least for UI issues.  Here in Saarbrücken,
>there is an on-going effort to create a GTK+ interface, and at the
>Belgian site of Louvain, Donatien Grolaux and Peter Vanroy have
>developed QTk which allows the creation of GUI's from descriptions: I
>love their approach and would like the user community to provide
>constructive critical feedback (of course, this goes for other
>libraries as well).

I have looked at Prototyper and it is very nice. I have not worked
with it to the extent that I can give any concrete feedback, however,
the performance on Windows was somewhat slow when using
several animated objects (sprites). This is an area where I have a
strong need/interest and thus my questions regarding use of COM
on Windows, and sockets vs. C++ extensions a while back. Which
approach is taken by QTk?

>As far as programming environment is concerned, I am interested in
>tools that facilitate the development, packaging, and installation of
>Mozart packages.  I have currently an alpha version of a new tool,
>called `ozmake' which is fairly intuitive and works the same on all
>platforms (by which I mean also Windows :-).

I read about ozmake in the documentation and it seems very useful.


>I don't know if there is interest in developing an IDE for Mozart.
>Personally, I am an old hand with Emacs and very comfortable and
>productive in that environment; but others might feel differently.

The development environment is an area which I personally feel is
very important for making people starting to use Oz. Let me
tell you how I got into using Oz. I work with programming tools for
users that are typically not programmes in a professional sense, for
instance children. Three years ago I came across ToonTalk, a
programming tool for children developed by Ken Kahn (see
http://www.toontalk.com). I learned that the computational model
of ToonTalk was based on concurrent constraint programming,
an area which I at that point had little knowledge of. I started to
learn more, for instance I read the introductory chapters in
Sverker Jansson's PhD on AKL. I was about to download and try
AKL when I learned from people at my department (the Computer
Science Department at Linköping University) that the current research
at SICS was focused on Oz. So I decided to try Oz. However, I
must admit that when I got to the download page an learned that
the programming environment was based on Emacs, I got kind
of "scared off" (even though I used FINE and Emacs extensively
15 years ago). I am just too used to IDE's like JBuilder and
VisualC++ (and before that Smalltalk and Interlisp-D) to be
willing to go back using Emacs. And I am not skilful enough that
I can customise Emacs. This alone was the factor that held me away
from trying out Oz. I read serveral (very good) papers on the
implementation of Oz and worked in parallel with an implementation
on a programming tool for children in Java, but it was not until Seif
Haridi came to Linköping and gave a demo of Oz that I actually
downloaded it and tried it out (because it was a really good demo!).

Using a development environment based on Emacs sends out signals
to people regarding what Oz is and what it is like to develop in Oz,
and certain people will be attracted by this, while others will shy
away. As I am an HCI-person, I am very focused on what the
communication goal is, and if the goal with Oz is to reach out to
many developers and researchers, my guess is that Emacs will
not do. Especially in the area of user interfaces. People like myself will
chose Squeak rather than Oz, even though Squeak is inferior in many
ways (and happens to have a very ugly default window colour scheme!).

There is something very attractive with the idea of a programming
environment being open and written in "its own" language (like
Interlisp and Smalltalk). It would be great to take such an approach
with Oz, and Prototyper seems as a good step in this direction.
And it would be great to involve HCI-people in Oz development
(perhaps this is already done) who could test the language and
the development tools with various categories of programmers.
I think that usability is at least as important as "computability"
to attract people's interest. I personally would love something
similar to Smalltalk (or even VisualBasic) based on Oz. Even
more intriguing would be something more radical, like ToonTalk
for Oz. This is along the lines of what I would like to do with Oz,
and if such a project would one day became reality, I would
love to participate if I would be given the opportunity. I have
also been thinking about starting such a project myself within
my work on children and programming.

Best, Micke






-
Please send submissions to users at mozart-oz.org
and administriva mail to users-request at mozart-oz.org.
The Mozart Oz web site is at http://www.mozart-oz.org/.