Take that writeup with a massive grain of salt. The author claims they spent 459 minutes on exercise 1.1, that exercise is this:
> Exercise 1.1: Below is a sequence of expressions. What is the result printed by the interpreter in response to each expression? Assume that the sequence is to be evaluated in the order in which it is presented.
There are then 12 simple expressions to evaluate. That is, it took them nearly 40 minutes for each expression.
Exercise 2.46 took them 535 minutes to implement. It wasn't even complex math, they needed to create a 2d-vector data type (their choice on implementation details) with a constructor, accessors, addition, subtraction, and scaling. That should not have taken 9 hours to complete (not by that point in the book at least).
In the first lecture, Abelson says Computer Science is neither a science nor is it really about computers. Considering the current ML paradigm, maybe CS has finally earned its name as a science.
The SICP video lectures with Gerald Sussman and Harold Abelson got me into Scheme and from there on Lisp. Although now I'm wondering if this would be better as a 'Show HN' submission.
i watched the lecture series during the pandemic and commented on many of the youtube videos. in at least one instance, a library function is used on the board that is not compatible with the current function signature in mit scheme.
I suppose it is something to do with the fact that it has been, what, almost 40 years since the lectures?
The fact that most of the code would still work is a miracle. That wouldn't work for, say, Java (which didn't exist in 1986). Nor C++. Nor Javascript (also not there back then). Fortran and C might be able to pull it off (but barely).
Remember, we didn't have computers worth the name back then. Shoot, we didn't even have dirt yet, just rocks.
If you are into SICP, you would probably like a nicely formatted html version of the book:
https://sarabander.github.io/sicp/html/index.xhtml#SEC_Conte...
And also this:
https://eli.thegreenplace.net/tag/sicp
The nicely formatted SICP is also available in downloadable formats.
EPUB - https://github.com/sarabander/sicp-epub/blob/master/sicp.epu...
PDF - https://github.com/sarabander/sicp-pdf/raw/master/sicp.pdf
Moreover, you can have SICP inside emacs by just downloading a package from Melpa:
https://melpa.org/#/sicp
There's also this interesting study about the difficulty and time requirement of SICP's exercises:
https://lockywolf.wordpress.com/2021/02/08/solving-sicp/
The math stuff is brutal.
Take that writeup with a massive grain of salt. The author claims they spent 459 minutes on exercise 1.1, that exercise is this:
> Exercise 1.1: Below is a sequence of expressions. What is the result printed by the interpreter in response to each expression? Assume that the sequence is to be evaluated in the order in which it is presented.
There are then 12 simple expressions to evaluate. That is, it took them nearly 40 minutes for each expression.
Exercise 2.46 took them 535 minutes to implement. It wasn't even complex math, they needed to create a 2d-vector data type (their choice on implementation details) with a constructor, accessors, addition, subtraction, and scaling. That should not have taken 9 hours to complete (not by that point in the book at least).
In the first lecture, Abelson says Computer Science is neither a science nor is it really about computers. Considering the current ML paradigm, maybe CS has finally earned its name as a science.
It is one of the most memorable first lectures in the history of Computer Science.
What about the current ML paradigm makes it a science?
Observing and testing phenomena we don't understand.
I guess it's been progressing from being math, to natural science, to social science
We have “laws” and routinely conduct “experiments” which are kind of unheard of in CS.
quite the opposite
The SICP video lectures with Gerald Sussman and Harold Abelson got me into Scheme and from there on Lisp. Although now I'm wondering if this would be better as a 'Show HN' submission.
Most of the code from the book is also available here: https://news.ycombinator.com/item?id=13918465
This is such a fun class!
i watched the lecture series during the pandemic and commented on many of the youtube videos. in at least one instance, a library function is used on the board that is not compatible with the current function signature in mit scheme.
Oh no.
I suppose it is something to do with the fact that it has been, what, almost 40 years since the lectures?
The fact that most of the code would still work is a miracle. That wouldn't work for, say, Java (which didn't exist in 1986). Nor C++. Nor Javascript (also not there back then). Fortran and C might be able to pull it off (but barely).
Remember, we didn't have computers worth the name back then. Shoot, we didn't even have dirt yet, just rocks.
> That wouldn't work for, say, Java
The ~29 years deprecated java.util.Date* methods would like to have a word. ;-)
*https://docs.oracle.com/en/java/javase/25/docs/api/java.base...
Which function?
would be better if you could just use AI to re-do those particular scenes in the video series..
I'd been hoping to do just this, but don't quite have the resources.
why can't you?
[dead]