I had an idea previously regarding the self-filtering of complexity.
Dolphin, for example, has an ANSI filter that only shows methods
that are ANSI compliant, which I find useful for quickly determining
the essence of a class. The problem with all these filters is that
someone has to do the division:

complexity level #one will include these methods, #two will include ...
etc, which is a real pain.

However, the system is smart enough to know for itself which methods
are most important. You could use #sendersOf to determine how
popular (important?) any particular method was. What I am suggesting
is that the browser could be set to only examine the 5 most frequently
used methods of a class, then 10, 15 etc. That way you could burrow
down to any level of complexity.

Would this be a useful addition to the present method categories for
beginners?

Peter


-----Original Message-----
From: Dwight Hughes <dwighth at ipa.net>
To: squeak at cs.uiuc.edu <squeak at cs.uiuc.edu>
Date: Wednesday, June 09, 1999 1:13 PM
Subject: System organization as "documentation" [was Re: Documentation
(...)]


>One of the most powerful, general, and evolving sources of information
>in Squeak is the system organization itself - or more precisely, the
>system meta-organization. The system categories and method protocols are
>a very powerful and pervasive form of "documentation". To this could be
>added further levels of meta-organization. For example: a general
>public(by default)/private classification for method selectors would add
>a significant level of "noise filtering" - allowing the browsers to be
>able to show only the selectors intended for general use without
>distractions (if desired). I see this being a level of meta-info about
>the method just like its protocol -- I'm not interested in the system
>actually enforcing private methods (yuck). You could certainly have more
>than just these two levels (and protocols could likewise be tagged
>public/private/whatever). A similar scheme could be applied to classes
>and even to system categories - in this case they might be called
>user/system categories, allowing another level of filtering if desired.
>Each of these would also add another level of reflective info for the
>system to use.
>
>The fundamental goal of all of this is to impart information in a highly
>leveraged fashion while greatly increasing the signal to noise ratio --
>you can find what you need to see when you go looking. The immense level
>of detail presented in a full-scale Smalltalk like Squeak, combined with
>no signposts that point to where to begin (especially if there are no
>wizards at hand to help), can make Squeak intimidating to newcomers. I
>think the real goal of the documentation effort (after plugging a few of
>the more obvious holes) should be to provide the knowledge of where and
>how to begin the journey, what you need to pack, and what fun things you
>can then do and see along the way. Great wads of detailed static
>documentation won't serve a highly dynamic environment like this very
>well (or for very long).
>
>-- Dwight
>
>
>Edward P Luwish wrote:
>>
>> I do not want this to become something like the discussion of "a message
is just a subroutine".
>> Not only is Smalltalk different from everything else, but the different
way one needs to think is the essence of its expressiveness.  Delving
through the corpus of code not only suggests how to solve a given problem,
it suggests new problems to solve.
>>
>> The question is how do we best communicate this, so that our community
grows larger and more influential in the world at large.  We have a terrific
opportunity, since this project is not only at the forefront of object
technology, but also of open source.
>>
>> There are experts out there who have some answers.  They have documented
why some practices may be better than others, and have discovered some
effective educational methods.  I will step back while they offer some
suggestions, and as a community we should expand on them, debate them, and
come up with a plan for a documentation project.  I will help in any way I
can.
>
>