379Re: [sicp-vsg] Is it active?
- Mar 31, 2007I started on SICP pretty much at the same time this mailing list went to sleep, and got answers to my questions on #scheme on Freenode (a fair share of thanks go to pjb for answering some of my questions). Regarding what Mr. Ansteth said, it has to be said that SICP is not for everyone. It is tough going, and some background in rigorous thinking is probably required (whether it's thinking about programming, or a background in some beginning university-level maths). It's not a question of "oh, I'm not smart as MIT students". You probably do not have the required background, or your background is rusty. And that's fine. You could either get the background and then go back to it, or work through the book How to Design Programs instead.
-GhalibOn 28/03/07, zarchne@... <zarchne@... > wrote:
On Tue, Mar 27, 2007 at 10:34:11AM -0700, Philip Ansteth wrote:
> prabhat137 <prabhat137@...> wrote:
>> I've just recently begun to study SICP on my own, and joined this
>> group looking forward to an active study group, which in absence of a
>> course or instructor, I could use as a forum for communication and
I'd like to take this opportunity to answer the roll call.
I got away from and behind the rest of the group probably at
the same time it was in decline anyway, but I haven't lost
interest altogether and would be glad of the motivation to
discuss and advance.
> A friendly warning:
> Over the last several years, I've made several attempts
> at understanding the Structure and Interpretation of
> Computer Programs. I was only partially successful.
> A lot of people say SICP is a great book. I tend to
> disagree. You'll want to make up your own mind, I'm
Can hardly not be true.
> Just to make my warning concrete, consider the exercise
> about Ackermann's function.
> I submit that the problem
> cannot be understood nor worked on successfully within
> the context of SICP alone. You have to go to some other
I disagree. The first three parts of the question are
completely straightforward and the last three parts require
only some observation and thought about the process.
> It may or may not be worthwhile to read up on Ackermann's function.
(Of course I am in the camp that it is unnecessary and, as
it happens, unhelpful.)
> But the point is that SICP is just no help. No
> explanations. No footnotes. No guidance.
(For none is really needed.)
> Can you skip the exercise if it doesn't interest you?
> Or is understanding it crucial to understanding later
> chapters? The authors just don't say. You're on your
I would say that the exercises are intended to cement your
understanding of the section to which they are attached, not
teach new ideas.
> I suspect that SICP was written to serve the pedagogical
> purposes of MIT's computer science department. I'm
> dubious about the practicality of studying it outside of
> that specialized context.
I say instead that it merely requires a certain degree of
mathematical maturity. (Exercise 1.13, for example, assumes
you already know how to do a proof by induction.) Perhaps a
rating system similar to Knuth's _Art of Computer
Programming_ (of which I have worked a far smaller
percentage of the exercises) would indeed be helpful.
I think, if you enjoyed (say, much of) your first year
calculus class, this is likely a good book for you. In
other words, it's not necessary to be an MIT student to
benefit from the book, only to be able to think like one.
In any case, I'm sure those of us still subscribed would be
happy to discuss and help out.
- << Previous post in topic