CSCMIC VS STklos

Hi, I graduated with Chemical engineering and left a dead end job late last year in oil sector. I kind of dabbled in programming on and off for the 7 years before that, learning bits and pieces of C, swift, iOS development, C++, machine learning, even algorithms and data structures; but nothing really “took off”. I used to mainly code simple engineering calculators. But there’s only so many engineering calculators the world needs and so since December last year I started studying CS “properly” - I did SICP and I learnt the basics of interpreters, assembly and compilers. I wanted to gain more experience with C before I did systems so I wrote generic data structures in C, even implemented a subset of Scheme in C .

Next I tried to study parsing from his book but didn't find parsing to be as enjoyable. I then went to the dragon book , which I had found to be very dense the first time I flipped through it an year ago. This time I found it to be much more approachable, thanks again to what I had read in Nystrom's book. Precedence and associativity made more sense when I read it a second time. But it turns out that Scheme's grammar is really simple: Expr -> atom | '(' Expr ')', so I stopped reading and started coding.

It's barely usable because it has no GC, and I may have to rewrite it from scratch for implementing one. In the meantime I am thinking of doing a rewrite in Java, and then do part 3 from Crafting Interpreters, where the author implements everything in C.

Still very much a newbie, I welcome criticisms and directions for moving it forward.

[1] https://github.com/TectonicFury/CSCMIC

STklos was indeed born in 2001, but it was the successor of Stk, which is much older. Its first release was in 1993. As far as I know, STkos did use code from STk, so I'm not sure one couldn't say it was born in 1993, as "STk".

That said, I'd just like to mention that I have made an effort to write internals documentation ( general: "hacking", bytecode: "vm") to STklos. It's an interesting implementation because it's easy to add primitives and change the compiler. The internals of the VM is actually a bit more complex... But it's really interesting code.

I've been working on enhancing STklos' macro system. The idea is to implement `syntax-case` properly and use it to obtain a better implementation of `syntax-rules` (the one in STklos is quite old and has some issues).

SRFI 143 has already been accepted, and there are recent pull requests for 170, 215, 217! :)