Roots of Lisp II

Since I wrote the post about Paul Graham's article Roots of Lisp about two weeks ago I've been searching for similar efforts to read, learn from and basically - steal. And I've got to say. There are some pretty impressive stuff out there.

Ur-Scheme is a self-hosting partial implementation of R5RS that runs on Intel x86 Linux.

The 90 minute Scheme to C compiler which is a presentation of how to write your own Scheme compiler (it takes 90 minutes to present, not to write the compiler).

The icbins Lisp to C compiler and interpreter in 10 pages of C by Darius Bacon. He's written an even smaller one (6 pages) but which doesn't include the interpreter.

Out of the three I suspect that icbins is what comes closest to answering my question (what kind of boiler-plate code would be required to support such an implementation) though I believe that it is written to support more (probably much more) than what is described in Graham's article or in the Lisp 1.5 Programmer's Manual.

More books arrived today!

When I started reading The Reasoned Schemer I realised that I had skipped The Seasoned Schemer. Not that that's not allowed or anything but I really felt that I needed to complete the series with it.

When I ordered it I also decided to give myself The Second Book of Go. I'm still rubbish at it though.

Got some more books today

I was plenty pleased today when I finally got the books I had ordered. I'm a huge fan of Donald Knuth and his book Things a Computer Scientist Rarely Talks About looks very promising indeed. I've already started reading it even though I had decided to start with The Reasoned Schemer which is a book about logic programming using Scheme. I can't say I *love* the style in which it's written (I was quite surprised when I first read through The Little Schemer a couple of months ago) but I can live with it ;-)

The book is using miniKanren (which I will download and install during the weekend). It will be interesting to try out some of the things they're describing in the book and I'm quite interested in the implementation aspects as well. It sort of touches quite a lot of things that I'm currently reading about or playing around with (logic object space in Pypy, implementing backward chaining in pyRete, etc)

I've been putting off the idea of spending some serious time trying to learn Prolog (because it looks and feels akward) but I'm not sure how much longer I can keep dodging it. It keeps popping up everywhere I look.