SICP Study Group

How do you measure learning?

Quote from "The Lifelong Learner":

Accomplishments, capabilities, social effects, strengthened sensitivities and convictions - in such consequences is free learning properly measured. Appraising such outcomes is, in a way, more difficult than taking a test at the end of a course and getting a grade. But the rewards are worth the challenge. And this kind of measuring encourages and aids further learning."

I would go even further: his list - Accomplishments, capabilities, social effects, strengthened sensitivities and convictions - doesn't just define a measure of learning, but also its purpose. To effect change in yourself, and in the world is (at least potentially) remarkable, worthy, and of value far beyond learning.

Progress report

We've just finished reading Chapter 1 of SICP, and have done a large subset of the exercises. We skipped some of the more math-intensive problems, although it was cool to implement integration and differentiation. It's possible to jump easily between the Scheme code and the mathematical definition, something that wasn't true at all of the Fortran and Pascal code we wrote in Numerical Analysis. Scheme seems well-suited to capturing mathematical ideas.

It's been extremely interesting to see how others approach the same problem. Some of the more substantial exercises have seen two or three different, valuable solution approaches. All of the details are in the chat transcripts and in the folders we've each created on the site to store our exercise answers. I think a 'best of' compilation that presented the exercises and their various solutions would be of value.

This is not the chat agenda for Week 1

I jotted down some thoughts in preparation for the first weekly chat, which I then proceeded to miss. It looks like it went pretty well, more thoughts later. Here, for posterity, are my thoughts at the end of week 1.

0) It might be helpful to read the 'Opening Question' pattern in the Knowledge Hydrant draft, at http://www.industriallogic.com/papers/khdraft.pdf

Here's the advice, excerpted from the pattern:

'Begin dialogues with an opening question or series of questions that penetrate into the heart of a reading, puzzling and/or challenging participants, and leading a group to seach for a work's most profound ideas. Let people volunteer to ask the opening question, and encourage them to record their questions to improve them overtime.'

1) What did you think of the idea that computer science isn't science or about computers? They used the phrase 'procedural epistemology', epistemology being 'The branch of philosophy that studies the nature of knowledge, its presuppositions and foundations, and its extent and validity.' (dictionary.com) Do you think of programming as a 'novel formal medium for expressing ideas about methodology'?

2) I'm thinking that the real due 'date' for the week's assignment is the start of the weekly chat. This is presuming that we'll discuss the assignment and solutions during the chat, and that this will serve as our grading/feedback mechanism. This may be assuming too much... but it's worth a chat.

3) Pascal's Triangle, the golden ratio... there's more math here than I usually run in to, and perhaps it's the same for you. My take on this is to try what we can, find pointers to fill in the gaps in our education and skills, and press on through it... but I was curious to see what everyone's mood was about this.

4) I had to go back and think over the differences between applicative and normal order evaluation, and write out for myself the sentence 'Instead, it would first substitute operand expressions for parameters until it obtained an expression involving only primitive operators, and would then perform the evaluation.' (page 16). Did anyone else wrestle with this? Can you create an example of an expression that evaluates differently based on the method used?

5) The preface described the programmer seeking 'perfection in part and adequacy of collection'. it got me thinking about the typical parts I use when writing code - variable, function, loop, array, etc... and it got me thinking that I'm still writing Fortran, even though I work mostly in OO or OO-based langauges and have for some time. It made me think about the language I use when coding, and wondering if I can make it richer. So, what collection will suit? And what's perfect?

Session 1

I started reading last night, and fired up DrScheme at lunch today.

It's funny, even though I understand the idea of prefix, infix and postfix, and though I've read the examples, it's very easy to type (a + b) rather than the proper (+ a b). I wonder how many other Fortran/Basic/C habits I will wrestle with.

The Learning Log

From 'The Lifelong Learner' :

The most important tool of free learning is a log-or journal, diary, notebook, whatever you choose to call it. Always have a pad and pencil on hand for jotting down ideas, thoughts, feelings, and even dreams. ... The purpose of a log is to help you keep track of those aspects of your learning and growth that seem most exciting. Express your ideas, insights, reactions, reflections, problems and doubts. Also use the log to record your reading, viewing, listening and interviewing. Later on it may be a useful place to project your plans, dreams, aspirations and experiments. ... Through this basic tool of free learning you will become increasingly conscious of what your daily experiences mean. You will be better able to see patterns in your growth that you might otherwise miss. You will heighten your awareness of your own learning process-what works for you and what doesn't and why. When you hit a plateau, reviewing the past month's journal entries will help refocus your self-education most effectively.

Learning a new language

"Learn at least one new [programming] language every year. Different languages solve the same problems in different ways. By learning several different approaches, you can help broaden your thinking and avoid getting stuck in a rut." --- The Pragmatic Programmer

After a long time, after Fortran, PL/I, 360 assembler, Basic, Pascal, C, 8088 assembler, awk, C++, Visual Basic, SQL, Perl, Python, PHP, I come to Scheme. Long admired, long wondered about, long mystified by, we are beginning a study of 'The Structure and Interpretation of Computer Programs', in which we will learn Scheme, and so, in a manner of speaking, we will Lisp.