I started doing something similar a couple weeks back -- trying to recreate the SICP LaTeX files so that I could generate my own beautiful PDFs, instead of using the slightly wonky HTML edition.
I even sent an email to Hal Abelson asking after the original TeX files, figuring it was a shot in the dark. I was blown away when he sent me a response... in less than an hour! He told me that he didn't think they even had a copy of it anymore, and that he didn't think it would be very useful in any case, but it was still really cool to get an email from a programming legend who has touched so many lives.
I wonder what format MIT Press has the book in - if it's a ps file or similar it should be possible to at least convert that into a PDF that looks exactly like the printed book.
True that would make it easy to produce a PDF but ePub is more flexible (and enjoyable) on the iPad than a PDF. Which is why the original author (from GitHub) created it.
It's probably worth running the HTML through tidy to get cleaner XHTML and then spending a little while trying to get http://code.google.com/p/epubcheck/ to be a bit happier. In general, EPUBs that "pass" epubcheck have a better chance of working interoperably in a lot of EPUB readers.
I pointed this out in another SICP thread, but it's worth restating for anyone who might be interested in Clojure but not want to invest in Scheme for SICP: The subset of Scheme that SICP uses is so small and simple, it doesn't take much effort to use that to do the exercises in the books. Doing everything in Clojure would likely distract from the core concepts, many of which are related to Clojure's philosophies.
I would be sad if anyone waited for one of these SICP-in-Clojure projects to be finished before even starting to read SICP.
It looks great; thank you! It will be nice to be able to put this on my Kindle and have it next to my MacBook as I work through examples. Flipping between the book and dev environment on a laptop screen is not as good.
Would be interested in getting this into my kindle app on my android phone, too. Found some instructions about converting to mobi format, but not sure how to sync to Android Kindle then.
I connected my Android phone to my Ubuntu PC using the USB cable. The phone's SD card was mounted as /media/XXXX-XXXX folder, which has a kindle subfolder. I copied sicp.mobi to the folder. That's it.
FBReader (http://www.fbreader.org/) is an open source ebook viewer for Android with wider format support (though it doesn't do PDF), greater configurability and a somewhat more no-nonsense UI approach. A particular highlight is the brilliant integration with ColorDict and other dictionary apps.
A little off topic: What's a good way to convert epub (or other ebook format) into HTML? Not all in one page like pdftohtml does but as tree nodes and a table of contents and all that.
Slightly off topic but I do not want to post another Ask HN question: what implementation of Scheme would you guys recommend/use? Working smoothly would emacs would be appreciated
MIT Scheme (http://www.gnu.org/software/mit-scheme/) includes a version of EMACS that runs in the same Scheme address space and is therefore well integrated. It's entirely sufficient for doing SICP.
Further update: 6.184 is being taught by the best of the last batch of 6.001 TAs, some of whom taught it 6 or 7 times. Their recommendation of Racket is therefore likely to be well informed.
I started doing something similar a couple weeks back -- trying to recreate the SICP LaTeX files so that I could generate my own beautiful PDFs, instead of using the slightly wonky HTML edition.
I even sent an email to Hal Abelson asking after the original TeX files, figuring it was a shot in the dark. I was blown away when he sent me a response... in less than an hour! He told me that he didn't think they even had a copy of it anymore, and that he didn't think it would be very useful in any case, but it was still really cool to get an email from a programming legend who has touched so many lives.
Too bad I forgot to ask for an autograph! ;)