Hacker News new | past | comments | ask | show | jobs | submit login

Every language should be faster* than C in some way otherwise no-one would use it.

Scripting languages are faster to code in.

Fortran is faster at complex matrix maths.

Lisp is faster to code in and faster to run for a large set of computing problems.

C wins the trade-of of speed-to-create vs speed-of-execution for most system programming tasks but usually when compared to bare assembly code. But there are a large number of problems that it is horribly unsuited for. Hence we end up with c-programs with accidentally embedded lisp implementations.

* - where faster can also mean faster to code to a secure standard.




Correct me if I am wrong, but somewhere I read that a while back, Lisp implementations caught up to and surpassed Fortran implementations in terms of math speed... suggesting that Fortran's main strength was finally outdone.


Richard Fateman published two papers on this, one where MacLisp beat out Fortran, in 1973 (http://portal.acm.org/citation.cfm?doid=1086803.1086804), and a really comprehensive paper about Common Lisp in 1995 (http://www.scribd.com/doc/49112889/fateman-et-al-fast-floati...).


> Hence we end up with c-programs with accidentally embedded lisp implementations.

See Greenspun's Tenth Rule: Any sufficiently complicated C or Fortran program contains an ad hoc, informally-specified, bug-ridden, slow implementation of half of Common Lisp.

And of course the famous corollary: Including Common Lisp.

http://en.wikipedia.org/wiki/Greenspun%27s_Tenth_Rule


XML is proof that Greenspun's Tenth Rule applies to data, as well. :)


I've said this before, but JSON is actually more like sexprs than XML is.


Agreed, but XML does a far better job of re-implementing sexprs badly.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: