A different view of Smalltalk

Les Tyrrell tyrrell at canis.uiuc.edu
Tue Aug 17 16:04:47 UTC 1999


We seem to think alike- you've been doing many of the things that I had been thinking
about doing.

Check out my Swikis:
    http://oasis.canis.uiuc.edu:8080/Squeak.1
        ( especially http://oasis.canis.uiuc.edu:8080/Squeak.54 )
    http://oasis.canis.uiuc.edu:8080/Oasis.1


The notation that you have here is very similar to the one that I developed for the
results of the code
analysis that is done by what I now call "Oasis".  The only differences is that I was
still in a bit of
a quandry about how best to deal with multi-argument messages, and I depicted a message
send
with a little arrow underlining the message, so that for instance "self size" would have
an arrow
from size's box, running underneath size, pointing at self's box.  I only used this on
paper, as Morphic
didn't exist back then.  I have a UI framework that I was working on to enable this sort
of thing, but
it's still very undeveloped.

I used this notation as a summary of the objects that appear ( everything in a box is an
object,
and now it's easy to say that and have people understand what you mean ) and the
expectations
that we have on those objects.  Finding those expectations, and then making sense out of
them
is one of the core tasks that Oasis's analysis system does.  Showing those results in
sensible
fashion has been hard, not so much due to the notation ( which I think is fairly clear )
but due
to the enormous amount of information that gets extracted, way beyond an order of
magnitude
greater than the original code size.

In short, I think it is a very nice contribution, one that experts and beginners can make
good use of.
I still have trouble explaining what it is that Oasis is doing to people who are quite
advanced,
but if I could point at just one method outlined as you have done it is very clear what I
mean,
and why I have bothered putting so much effort into Oasis.

As for being a writing tool, I'm not so sure- just have to try and see it can be done.
BTW,
as I'm sure you are probably aware, there have been other graphically executed syntactical
depictions for Smalltalk, the most recent being Golgi and there being an earlier one for
VisualWorks
that was quite nice.  I think that only Golgi actually allowed editing that way.

thanks!

-les


----- Original Message -----
From: Bob Arning <arning at charm.net>
To: <squeak at cs.uiuc.edu>
Sent: Tuesday, August 17, 1999 8:39 AM
Subject: A different view of Smalltalk


> Recent discussions started me thinking about different ways to display, understand and
manipulate Smalltalk code. SyntaxMorph is an experiment that provides a graphical
representation of Smalltalk syntax with popup menus to identify, examine and browse
individual features. A SyntaxMorph view of a method can be created by selecting "BOBS
SYNTAX" from the browser message list menu.
>
> Questions to answer:
> 1. Can this provide useful information to beginning Smalltalkers? Intermediate?
Advanced?
> 2. Can this become a tool for writing code as well as reading it?
> 3. Is it just a lot of pretty pictures?
>
> Cheers,
> Bob
>
> SyntaxMorph (for 2.5) can be found at:
>
> http://www.charm.net/~arning/SyntaxMorph.17Aug933am.cs
>





More information about the Squeak-dev mailing list