As some of you know, I've been working for a long time on a replacement interpreter for Joy in Chicken Scheme. I chose Chicken because its compiler generates...
use the extension, conform to the spec. ... maintain...
sa@...
Mar 23, 2006 8:31 pm
2783
... Alas, it's not so simple. The Joy papers explicitly say that the maximum size of a set is implementation-dependent, so 31-bit sets are in fact conformant....
... Urk. Not to mention that giving MORE bits would make programs written for Chicken-Joy not compatible with C-Joy. I'm in favor of unlimited bits, though,...
These are difficult questions. I seem to remember about 20 years ago there was a lot of discussion about the advantages of tagged architectures, in which ever...
... Of course, that would be tough in the Unicode world, with a theoretical maximum of 17 * 2^16 = 1,114,112 characters! ... Last night it occurred to me that...
... is ... the ... The problem isn't with "cons", it's with "uncons" and friends. The Joy language doesn't distinguish between lists and functions, which is ...
Can someone (probably Manfred) explain to me how modules and hiding work in Joy0/Joy1? That was a part of the source I didn't need to figure out in revising...
... First, some general comments. The symbol table and its associated functions could be implemented in several ways. The choice would depend much on what the...
... Ah, this was what I didn't grasp before: the symbols *include* their names, rather than being pointed to by their names. In fact, the symbol "bar" as used...
I have written comments about the implementation of HIDE and MODULEs close to the functions inside main.c. I wrote them using the vi editor using the terminal...
... Unfortunately, Yahoo Groups stripped the attachment, even though it is a plain-text one. If you email it directly to me, I can resend to the group without...
Manfred sent these to me by private mail. Here they are: /* file: main.c-COMMENTS */ I comment on the functions in main.c which are relevant to the symbol...
an expression in the combinators S, K, I: Ix -> x Kxy -> x Sxyz -> xz(yz) can be optimized (shortened) by systematically replacing certain patterns with the...
stevan apter
sa@...
Apr 22, 2006 12:54 pm
2795
... This is just from memory: Turner introduced some variants of the classical combinators by others which do what in concatenative terminology might be...
Sorry for the OTness but I know there are some very clever folk on this list who may also be involved in NP stuff and tackling them with domain specific...
If you asked this 20 years ago I would have no reservations recommending Scheme. It has always been a good language for creating domain-specific languages. ...
chapter 16 of p-j has what i'm looking for, and googling "director strings" turns up some interesting research material. thanks manfred. ... From: Manfred Von...
stevan apter
sa@...
Apr 26, 2006 10:56 am
2799
hi martin - welcome back. i'm a little puzzled. why should NPness matter in the choice of language? i can understand why one might care (or not) about...
sa@...
Apr 26, 2006 3:37 pm
2800
... Sounds like you need a language that you know well enough that it doesn't get in your way. C++ is never that language (and I know it pretty well). Python...
or pick a language which has a subcommunity actively involved in the application domain. so: lisp, scheme, or one of the functional languages. ... From:...
stevan apter
sa@...
Apr 28, 2006 1:52 am
2802
I was wondering why the term "concatenative" was chosen over the term "compositional"? The semantics of the Joy language if oftern explained as function...
charles peirce (the inventor of pragmatism) renamed his doctrine "pragmaticism" after william james started using the original. the new term, peirce thought,...
sa@...
May 2, 2006 4:06 pm
2804
... I've been blamed for that choice. I think it makes sense to me in that "concatenative" describes an elementary property of the language type that describes...
... Looks good to me. Have you played around with/read the documentation about strongForth, at http://home.vrweb.de/stephan.becher/forth/index.htm? Your work...
I notice in stack based languages certain symmetric programs reduce to no-ops: f1 = [swap swap] = [] f2 = [dup pop] = [] f3 = [cons uncons] = [] f4 = [dup swap...
... Well, by definition if a word has a "symmetric" counterpart -- a mirror-image reversal -- then following the word with its mirror image would be a no-op....
our own brent kerby and hilton campbell: befreak. http://tunes.org/~iepos/befreak.html and a k implementation with GUI: http://www.nsl.com/papers/befreak.htm ...
stevan apter
sa@...
May 11, 2006 10:31 pm
2810
Sort of OT, but this reminds me of a peep-hole optimizer I once implemented for a sliding block puzzle solver: two adjacent moves that move the same piece into...
