Hacker News Reader

Comment by silentbicycle | Hacker News Reader

Well, the benefit of writing a Lisp-ish interpreter in Lisp is that you can focus on the semantics, rather than how things like garbage collection, type tagging, and symbol interning are implemented. When using another language, the relevant infrastructure might not already be lying around, but you can take advantage its novel features instead.

View on HN

That may be valuable, but there's value in this too.

I've had a play this morning, and am already most of the way towards having the readline sitting in twisted and the instance feeding a webserver. Which is to say - lisp notation with all the benefit of python's great libraries, and I understand the layers completely.

No kittens were killed.

I wasn't the one who said the trite thing about kittens - I think writing toy interpreters and compilers is a great way to figure out how languages work.

yeah ... that was my point ... it's a tradition to write Lisp in Lisp / Scheme in Scheme but it's also a testimony to the simplicity of Lisp 1.0. The actual first functioning Lisp interpreter was written in IBM 704 assembler, but once that was done, everything else from that day forward could have been done in Lisp. A fair amount of it was, once macros and Lisp compilers showed up.

Lisp is not particularly unique in that regard. Forth is even simpler (perhaps too simple), and it's not that hard to write a naive implementation of Joy or Tcl.

Besides, piggybacking off the existing Lisp implementation doesn't help much when you're trying to figure out how the infrastructure (GC, etc.) is implemented. I learned a LOT implementing a Scheme interpreter in OCaml a couple years ago, and wouldn't discourage anyone from doing so in whatever language they feel most comfortable with. (I'm actually taking a break from writing a small compiler right now, for a proprietary database query language we need to convert.)

I think you are/were mostly being voted down because of your tone.

Reply on news.ycombinator.com