[olavk]

I worked through the book in my spare time some years ago, and yes it probably took me hundreds of hours. I don't know the exact amount of time, but it was definitely in that range. I did not have any CS or math background, but a lot of practical programming experience. I was absolutely worth it.

I did not watch any of the lectures, just worked through the book, so I can't say if watching the lectures will make it easier or just take more time.

[dajomu]

Thanks, good to know how much time I need to apportion.

I'm trying to find a way that I can fit it around work/family schedules and not let my concentration go too stale. I'm thinking of trying for an hour a day, but that feels so prescriptive...

[olavk]

The only advice I can give is, do not set goals like a chapter a week or n exercises a day. Sometimes a section or exercise is just hard and you have to take the time necessary.

[Jach]

This might be considered heresy but the first two chapters can go very quickly, and taken alone can be very impacting, so I'd recommend just them as an initial goal. Once you accomplish it (sooner than you might think!) you ought to use the momentum to continue on, just know it's a steep book for the last few chapters...

A quote I keep in mind: "The reader who has read the book but cannot do the exercises has learned nothing." -- J.J. Sakurai

I'm not quite as strict as Sakurai suggests -- sometimes it's helpful for me to plow ahead anyway and try to come back to the exercises later... Sometimes it's never, but if I do some exercises, integrate some of the non-exercises with other bits of knowledge, I still feel like I've learned something rather than nothing. "Gotten familiar with" might be the best phrase, and there's a nice technique Mike Acton made an infographic of for getting familiar with a stack of books on short notice: https://itsyourturnblog.com/you-cant-cheat-at-learning-quick...

[dempseye]

You say it was absolutely worth it. What did you gain from the experience?

[olavk]

My background is as a self-thought programmer. Never had any formal education related to software, but have worked professionally as a developer for many years. This gives some solid pragmatic experience, but the "fundamentals" tend to lack. I considered a compiler a form of black magic which I had no hope of ever understanding.

SICP gave me a throughout tour of basically all layers of the stack. Understanding the fundamentals of compilers just makes it easier to understand the design decisions behind individual languages, which makes them easier to learn and easier to leverage. SICP is also a journey throughout various paradigms - imperative, functional, stream based, lazy, OO, constraints solving etc. Having a solid grasp of various computing paradigms just give you much more freedom in your approach to solve problems. It also helped dispell any religion I had about particular platforms and languages - I realized religion is just comfort + fear. Experimenting with things like continuations and backtracking was almost mind-altering for me, since I thought the stack was an immutable law of nature and just not a particular design.