Hacker News new | past | comments | ask | show | jobs | submit login
Peter Norvig's review of SICP (amazon.com)
41 points by andreyf on Jan 31, 2009 | hide | past | web | favorite | 18 comments



My brother is starting to learn a bit of programming, and was looking for books to read. I suggested SICP, so he went to Amazon and immediately picked up on the fact that almost all of the review scores were 1 or 5.

Interestingly, he didn't bother reading the favourable reviews--just the unfavourable ones. His argument was that in reading the unfavourable reviews he could better pick up what the book's strengths were: what did it teach that many people weren't ready to learn?


I wish my peers were like your brother.


SICP really changed the way I look at computer science. So enlightening: data and commands are interchangeable etc. I was surprised to see so many bad reviews on amazon. Although flipping through them, it seems they just don't 'want' to understand the book :(


it seems they just don't 'want' to understand the book :(

I think it's more that some brains are wired for that kind of abstraction and some are not. Those who love the book are mostly the ones whose are.

A similar (if not the same) phenomenon is that there are far more mediocre programmers than good ones... a simple fact which determines nearly everything about the software industry.


some brains are wired for that kind of abstraction

In my experience if someone can't program it's not becuase they have the wrong static structure, it's that they're unwilling to dynamically metaprogram their thought processes and fundamentally reshape their mental models.

far more mediocre programmers than good ones

Same goes for actors and boxers and cops. This seems, almost tautologically, true of any industry. Why is this meme so prevalent?


In my experience if someone can't program it's not becuase they have the wrong static structure, it's that they're unwilling to dynamically metaprogram their thought processes and fundamentally reshape their mental models.

Either way, the point is the same: 90% of people (if not 99%) are averse to the kind of abstraction we're talking about. I have no idea how many are unable and how many merely unwilling; I'm not even sure the distinction matters or how you go about deciding it. (By the way, where did you get that phraseology? It sounds like NLP.)

Same goes for actors and boxers and cops. This seems, almost tautologically, true of any industry. Why is this meme so prevalent?

It's because the demand for programs vastly exceeds the supply of good programmers, which is the fundamental theorem of software economics (at least in our lifetime) and skews everything about the industry. Obviously the mediocre outnumber the competent at any activity, but that usually doesn't become the dominant fact about that activity. For example, there are many more bad musicians than good ones, but the demand for music doesn't vastly exceed the supply of good musicians. If it did, Lord knows what sorts of contraptions people would call instruments and mass-distribute. "We can't have pianos around here - how will we hire 100 musicians to play them in the future? And if guitars are so great, how come they're not more popular? Better to be a blubby-music-box shop like everybody else! Our clients expect it!"


the kind of abstraction we're talking about

You know, I wonder to what extent it's more the peripheral hassles than the actual abstractions. There's no way in hell I'd be a programmer if it meant working with punch cards; it'd be above my irritation threshold. There might be a lot of people out there who would enjoy programming, but installing and configuring a compiler and editor, for instance, are simply too frustrating/boring to them.

There weren't many photographers when you had to develop your own film. (I think I'm paraphrasing a Steve Yegge essay... I'd look it up if I weren't on my iPhone at a volleyball tournament :p )

where did you get that phraseology? Sounds like NLP.

That and general semantics I believe. I haven't read much of the primary works in either (yet), but picked up some of the concepts/terms from Robert Anton Wilson.


I recently came across SICP's amazon page and based on this and Paul Graham's reviews, I decided to buy it and work my way through it. It arrived yesterday and now I am about halfway through Chapter 1. So far its a pretty good review and I've gotten a much clearer understanding of some things.

I finished school almost 2 years ago, so a lot of things I found confusing and odd about Scheme when I first used it in school have had time to settle in my mind (lambdas, first-class functions, etc). Hopefully it will be worth my time.


I started working through my copy about a month ago and have no regrets about spending a lot of time with it (still have 1 chapter to go). While I might not have felt very enthusiastic about working through all the exercises, they're VERY helpful and a central part of the book.

I found that SICP helped hone my skills in finding elegant solutions to certain problems and improved my coding style. On the downside, I switched back to C++ tonight and found myself putting parenthesis in all the wrong places for the first hour. It's a small price to pay.

Enjoy!


Norvig is head of research at Google and the author of a leading book on AI.



Leading != canonical.


What's the canonical book? Norvig has, IIRC, written two widely-appreciated books on AI.


I wasn't clear. Norvig has written canonical books on AI, but they are not leading edge.


Why is this book so expensive? The cheapest used copy on amazon is ~$60.00...


Particularly when it is available online for free at http://mitpress.mit.edu/sicp/full-text/book/book.html


Particularly because it is available online for free.

For a normal book, you have cheapstake buyers, and then buyers who will pay a premium for a nicer book (hardcover, newer, etc). Now take the cheapstake buyers out of the market because they're all accessing it online for free. Economic magic does its thing, and the market price is now much higher.


The word is cheapskate not cheapstake.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: