<Darius>my opinion is that it's educational but not something i'm tempted to use directly in any other project so far <OriansJ>Darius: it is missing a FSF approved license <Darius>thought i stuck a license on the whole repo, but ok, thanks for the reminder <Darius>i started on some docs to check in, but am distracted by real-world problems right now <xentrac>I'm embarrassed to admit I still don't fully understand ichbins tho <xentrac>I need to go back and study it again <xentrac>Darius: that's a pretty harsh condemnation of Forth coming from someone who interned at Forth Inc. :) <Darius>the mid-80s were a different world :) <xentrac>well, maybe not of Forth, but of using Forth in real life I guess <Darius>it was a little bit like a more practical smalltalk in that older context <xentrac>like smalltalk because you had interactive development and hyperlinks to definitions? <xentrac>and untyped is sort of like dynamically typed I guess <Darius>interactive development, agile control over the environment, full source to it <xentrac>so what changed is that Smalltalk became practical? :) <Darius>and using that access & control was a normal thing to do in that culture <xentrac>people in Python get upset when you change the value of 5 <Darius>i think smalltalk/v had recently come out (or some earlier incarnation in a different name?) but really pushed the limits of the pc <Darius>i never had a chance to play with that, anyway <xentrac>yeah, and wasn't available for the PDP-10 <fossy>Darius: which varient of lisp? <xentrac>the name stands for "I can hardly believe it's not Scheme" <OriansJ>To reduce manually having to spot cases where M2-Planet will generate a binary which will segment fault because the C program initializes a local inside of a Loop; I'm simply going to have M2-Planet display an error message with that context information and halt. (This will result in some "working" programs breaking but will minimize the problem long term) <OriansJ>and the related mescc-tools fixes are now up <xentrac>hasn't been such an unproductive weekend after all :) <OriansJ>xentrac: only an hour of coding time is less than ideal <xentrac>my friend Jeff Barr told me that one of his most productive side project coding years was on Carmen's Headline Viewer <OriansJ>and a few minor improvements to mes-m2's macro support got squeezed in too. <xentrac>which he named after his wife, but mostly just called it HeadlineViewer at, I think, her request <xentrac>so I think that year he had six kids, one of which was a baby and one a toddler <xentrac>he said he wrote all of CHV in fifteen-minute pieces, which was possible in significant part because he never got up from the chair unless the code was working <xentrac>but that was the largest piece of uninterrupted time he could manage that year <OriansJ>writing a standards compliant Scheme in a C subset in 15minute increments. Sounds like a religous challenge. <xentrac>like doing a pilgrimage on your knees? <xentrac>well, HeadlineViewer also had some advantages there :) <OriansJ>xentrac: more like a pilgrimage on one's back <xentrac>sounds like a porn story about temple prostitutes <xentrac>not familiar. was that Joseph's brother? <siraben>At the point where you have self-compilation, it gets real meta real fast. Is my language the "real" one using C is an intermediate language or is C the "real" one? From an philosophical (?) viewpoint it's indistinguishable. <siraben>Speaking of meta things, is anyone aware of a full implementation of the Futamura projections? <Darius>siraben, that's not quite that -- i was taking the tack that it should be more straightforward to write the same thing you get by the third futamura projection, but directly <Darius>also i'm not sure that version is in a working state -- maybe only the old/ directory? <Darius>it did work once but it's actually much older than the 7 years ago it was put on github <Darius>that thesis did seem to bear out, at least for that simple first-order sublanguage <rain1>that's a good question.. where is some working code for futamura <Darius>rain1, siraben, i dunno if this exactly answers, but nada amin / tiark rompf on collapsing towers of interpreters sounded relevant and recent <Darius>and i know you can find the repo <terpri>labelled "Source Program and Results of Test Runs of Self Applicable PE" on futamara's homepage (https://fi.ftmr.info/); it looks like LISP 1.5 (with handwritten M-expressions in the margins!) <Darius>i definitely remember listings in the jones-sestoft partial evaluation book <rain1>it looks like it could be possible to get Similix working <siraben>Darius: Yeah I recall that talk. I haven't looked into the papers though. <siraben>Fortunately it looks like they have small-step and big-step operational semantics for the translation, however I wish it took a more abstract, perhaps even categorial, approach. <rain1>> (similix 'append1 (list '(1 2 3) '***) "append.sim") <rain1>front-end flow bt sp eod oc rl <rain1>((define (append1-0 l2_0) (cons 1 (cons 2 (cons 3 l2_0))))) ***pgreco_ is now known as pgreco