Last year I embarked on somwhat of a journey to find a better language for my home projects after getting a bit frustrated by python's lack of blocks and general cruftyness. After a couple of months of trying various different things I settled on Gambit Scheme for my spare-time data indexing project. A minimal core language with uniform syntax and macros. Lots of potentual for adapting and building language features that make sense to me. Sorted.
Then last week I got round to reading Richard Jones's minimal forth code. A language compiler and runtime in a couple of pages of well documented X86 assembler; small enough to read and understand the whole system in a bus journey.
Like Scheme, Forth has the ability to construct new language features, allowing user code to get between the parser and the evaluator to modify the language itself. Also, like Scheme, it has a uniform syntax - words separated by spaces, which makes re-writing code on the fly practical. Both these features mean that a fully fledged programming environment can be bootstrapped up from a minimal core. And the hook is: that minimal core is so much smaller for Forth than it is for Scheme.
One thing lead to another and I got interested in stack languages. Now I'm looking at Factor and wondering...
Factor has ticks in all the right boxes: minimal core, machine code compiler, macros, continuations, lightweight threads and message passing concurrency. It's also the tersest language I've seen - code to do something always seems a fraction of the size I expect it to be. On the other hand it's so radically different to anything else I've programmed in and it makes my head hurt. Could it be a contender to knock Scheme off the top spot?