[racket] Exploratory programming?

From: Richard Lawrence (richard.lawrence at berkeley.edu)
Date: Tue Nov 30 11:55:54 EST 2010

Dear Racketeers:

Apologies in advance for what will probably be a long-winded newbie
question.  I'm wondering how one goes about "exploring" the Racket
libraries.  That is, how does one find out things like:

1. which modules are likely to contain functions useful for a particular
task
2. what names a given module exports
3. more information about the objects exported by a module: e.g., 
brief descriptions, their types, what methods they respond to, ...

Here's some background. Most of the programming I have done is in
Python. One of my favorite features in Python is that

  >>> help(X) 

gives useful information for just about any object X. This is great for
exploring modules, learning about the methods and member variables of
classes, and finding out what functions do; it's an important
"bottom-up" tool for exploring the language and libraries. In Python, I
can often think, "I need something similar to, but not quite the same
as, X; X is defined in module M; so let's do help(M) to see if M has the
function I want." Often, I can tell at a glance over the docstrings that
the thing I need -- or something related to it -- is there.  (This kind
of "bottom-up" exploration is complemented by the documentation for the
Python standard library, which provides a nice "top-down" reference.)

I know that REPL-based interaction is de-emphasized in Racket.  Does
that mean that the kind of "bottom-up" exploration of the language that
I'm describing is de-emphasized as well?  Should the documentation be
the first (only?) place I go to learn about modules and their contents?
If so, what's the best way to access and search the documentation while
programming?

Please understand that I'm not ragging on Racket's documentation here;
the documentation is truly excellent.  I guess one way of asking my
question might be this.  I tend to think of the documentation as
something I must sit down and read for a while, partly because the
examples are often extended and build on each other, and partly because
of the psychology of jumping out of Emacs into a browser.[1] Is there a
different way of using and thinking about them?  In particular, is there
a way of using them that corresponds to the uses of help() in Python?

Thanks for your thoughts!

Best,
Richard


[1] I have so far mostly stayed away from DrRacket, because I know Emacs
well and haven't felt the desire or need so far to learn another
environment.  Is that my problem?  Does DrRacket support the kind of
bottom-up exploration in the documentation that I'm trying to describe?  



Posted on the users mailing list.