sbp: I guess it's going to try to grab from pgp.mit.edu?

sbp: I was perhaps disingenuous about ascii being the pertinent item. after all, as I say, I could do lexical scope without the ascii representation

sbp: asciilifeform: how did you find SEdit? I have only spoken to one friend who used the Interlisp-D machine, and I don't think he said anything about SEdit

sbp: mircea_popescu: no, but I can generate one. I was going to say "easily", but you know what software is like

sbp: greetings mircea_popescu! I am Sean B. Palmer, very pleased to meet you

sbp: fexprs as arguments to functions must be typed. that's it. that's all we need

sbp: well, we make one concession to make fexprs that we can compile

sbp: Wand's theory, and all the other pissing against fexprs from Pitman onwards, is based on the untyped lambda calculus. naturally. and lisp is untyped in this sense; the types come at runtime

sbp: yeah, well the answer to that is in itself trivial. a child could have come up with it

sbp: you remember that Mitchell Wand proved that the theory of fexprs is trivial? i.e. that fexprs cannot actually be compiled, and they must therefore be runtime components

sbp: I mean I figured out how to make them compilable

sbp: but perhaps my favourite thing was that I managed to fix fexprs

sbp: when you look at lambda calculus, you get most lacklustre computer scientists talking about alpha-renaming and all this stupid shit that gets in the way, but when you use de Bruijn notation that stuff disappears entirely. it's an epiphenomenon, one of Ptolemy's epicycles

sbp: this was all made a lot easier by using de Bruijn notation internally for variables by the way

sbp: I'm not sure I care awfully about the runtime constraint. ("two speeds" of a computer and all that)

sbp: the drawback is that because you depend on execution frames (or whatever) for the data, which is what allows the use of dynamic scope of course, this has to be done at runtime. lexical scope would usually be computed at compile time

sbp: what I realised was that when you couple scope to s-expressions in this way, it essentially becomes a system of runtime assertions in which you can model not only lexical AND dynamic scope—by choice!—but other kinds of hitherto unexplored scopes too

sbp: so I spent some time thinking about how scoping would work if *lexical* scope were bound to s-expressions and not the ascii representations of programs. because when you think about it, that's all that lexical scope is: it's an artefact of ascii representation, and I thought that perhaps this was not the lispy way

sbp: well SEdit allowed the editorial process, that protean forge, to work in like direct manner

sbp: a bit like how the SCHEME-79 chip worked. that executed the cons cells directly, as you put it

sbp: the idea was that you edited the cons cells directly. there was no intermediate ascii representation. in other words, there was no byte array buffer on which the editor acted; the editor acted directly on the s-expressions in the machine

sbp: there isn't much about it on the web now. I think I found a single PDF describing it in detail!

sbp: also, I don't know if you remember, but the Interlisp-D machine had a program called SEdit

sbp: but this got me thinking about Lisp bytecode, and whether the best bytecode for Lisp might be Forth, in essence

sbp: well, it was reverse polish at least, not that this is the most interesting feature of Forth

sbp: have you heard of Reverse Polish Lisp? it was a language for the HP-48 I think from 1987. the idea was that it was meant to combine some of the features of lisp, the high level stuff (as high level as they could squeeze into a late 1980s calculator) and the low level Mooreishness of Forth

sbp: only that, and to enquire about Phuctor. I have not contacted you again for precisely that reason: I abhor tedum too. I did get a copy of Kogge, and I did review it, and I did create various systems based on that. but none were to my liking so far

sbp: long time listener, second time caller, as they say

sbp: I'm the Alert Reader from Loper 1361. we don't know one another outside of that, sorry!

sbp: I may have to endure citizenship just to save you from the ignominy of periodic bot commands

sbp: the joke that most historians neglect to mention is that it didn't apply to slaves

sbp: trinque: you recall the redistribution of land by Lycurgus of Lacedaemon?

sbp: I'll forbear my Roman name for now, but perhaps the invocations will come to my fingers sooner or later

sbp: I apologise for not having a PGP presence, the baseline of citizenship

sbp: Nelson mentions that computer is a misnomer, and quoted Von Neumann as calling them all-purpose machines. I tracked that down potentially to https://www.cl.cam.ac.uk/~rdm34/burks.pdf

sbp: ("Max Allen and Ted Nelson discuss the future of computers (1979)")

sbp: here is some more joy from the Times That People [CD]are Not To Recall:

sbp: that's some foxy loving. thanks asciilifeform!