Hacker News new | ask | show | jobs
by digitailor 4809 days ago
As cool as an approach that Factor has, I cannot fathom why it is being developed for x86 for the life of me. (Besides claims of cross-platform native executable compilation.)

On ARM it would be interesting- or if it had a highly modular structure that would allow it to be "easily" ported to other microprocessors.

But on x86? What niche is it trying to serve? Is it a language-lover's-language? I guess what I'm asking is, why would I use a "high-level" Forth on x86? There has to be a reason, I'm just missing it.
2 comments
doublec 4809 days ago
Factor used to run on ARM [1]. Back then Windows Mobile only had 64MB of ram and it wasn't quite enough to do anything useful in Factor.

I used to run a video sharing website written in Factor [2], sadly the server crashed a few weeks ago and I haven't recovered the data from the disk yet.

Factor feels more like a Scheme than a Forth to me when programming.

[1] http://web.archive.org/web/20111117015025/http://www.bluishc... [2] http://web.archive.org/web/20130118204336/http://tinyvid.tv/

link
digitailor 4809 days ago
Woah thanks for that link, there could be some assembly worth salvaging there.

Currently working on an ARM forth-idea with some scheme-ideas. There's this nagging intuition I have that there's a symbiosis between Lisp atom/Forth word and S-expression/Forth dictionary.

I have to get something going pretty quickly so I can't get too pure about the idea, as I don't want to get bogged down in the necessary GC yet.

Thanks for your reply, it already made me more interested in Factor from an implementation standpoint, even if I wouldn't use it on x86.

link
colomon 4809 days ago
Isn't Factor (at some level of abstraction) basically Lisp/Scheme with more minimal syntax? It replaces prefix notation and parentheses with postfix notation and a stack. All the benefits Lisps get from minimal syntax, Factor gets too.
link
digitailor 4809 days ago
That's exactly what I'm thinking, but I haven't had time to stress-test the idea or find any precedent. And remember i'm talking about a ~36kilobtye forth, not this large-sized and resource-intensive kind of x86 forth that I don't understand the reason for.

Theoretically you would have the HLL incredibleness of a lisp with the low level genius of Forth. I asked some old timers if the idea made sense and they were silent in specifics but not discouraging. Which would indicate to me they think it is a huge learning opportunity for me one way or the other. Have you done any deeper analysis?

link
bun-neh 4807 days ago
They are working on reimplementing ARM support in the next release. It is in an incomplete state ATM.
link