★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz And fuck the invisible hand you're talking about. It's not that I disagree with what you wrote, I mean, the market only matters about money, and it was wrong to borrow these early Lisp processors.
We are crypto-anarchists and we research and develop concepts for political reasons and goals, for the people.
screwlisp
@stman
I realize that I'm a one-trick-pony, but I continue to think that if your own work veers close to lisp, you can directly line your ducks up with the MIT AI lab code and emails at the time / https://tumbleweed.nu/r/lm-3/uv/cadr.html now. @amszmidt's one branches from the last state of the MIT one; the other-language compilers written for later lisp machines were and are proprietary and hence lost.
If you wanted something different, I guess Wirth's Oberon is that.
@theruran @50htz
★ STMAN ★ 🏳️🌈
So maybe we could say the current research / brainstorming effort in this thread is an attempt to generalize better these Lisp Machines developed in the 80', but for other reasons than the ones they were invented at first sight.
Generalization efforts are always giving good results in terms of fundamental research and high level conceptual brainstorming in our researcher group, particularly when working on fighting complexity.
★ STMAN ★ 🏳️🌈
Here is a human readable pseudo-code attempt on a simple C program. It can be improved, it's not definitive. It's just a try.
Now we'd have to do the same for 2 or 3 other very different languages, and then take a pause to observe all these pseudo-codes to try to see what other "sub-models definitions'" would be needed to configure a universal engine so that it can interpret all these languages.
★ STMAN ★ 🏳️🌈
@screwtape IF you want an access to the shared pad this text is, just ask. It's on a cryptpad.
★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz So now we have to choose two or three other languages, to do the exact same manual pseudo-code generation as I did with C.
Any suggestion ?
• Go
• Rust
• Lisp
• Python
?
But see, for Go, we will be forced to handle GoRoutines natively, meaning the engine must allow be able to handle threads / processes concept natively. This aspect of Go language, and the necessary model information to
★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz So yes, one of these sub-model data would contain all the microcode of each pseudo-code instruction, but we are researching other sub-models that would complement the "microcode sub-model data" to configure other aspect of the engine to fit several languages, and not expressed in the microcode.
Microcoded engines are actually engines with a single model configuration, containing the microcode.
★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz But in order to have a more universal engine able to interpret more than one language, we are trying to see how we could split a single model (microcode) into more sub-models, so that some specific aspect of each language can be better configured, or more simply.
We have already explored this strategy of extracting sub-models from "common code" in the past, and it always reduces complexity.
★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz Your inputs, contributions and remarks are always very welcomed in our threads, you always contribute very positively. I personnaly love when you break into our threads because it always helps. I value a lot your knowledge, expertize, culture and your mindset, for you to know.
★ STMAN ★ 🏳️🌈
@screwtape @amszmidt @theruran @50htz I haven't read all the informations of links you have shared with us yet, but I'm going to do it. If you really think we could team up with other persons researching in similar directions, it's all good ! It's also what we're looking for !
★ STMAN ★ 🏳️🌈
But to reformulate again, what we're trying to do in this thread, is to find how to use as much sub-models as possible for configuring such universal engine allowing to run any languages pseudo-code "natively" without a preliminary complex ccmpilation process, and how these sub-models would look like, their format, what they would actually represent.
screwlisp
@stman
More context, in the later lisp-machines my understanding is that the microcoding of low level commands was actually done in lisp, as lisp, with lisp, even though the microcode itself was not lisp per se.
Lisp also has the low level microcoded bits in it that the high level bits would be made of inside of it https://www.lispworks.com/documentation/HyperSpec/Body/f_ldb.htm . I don't think that other languages are amenable to the interpretation we're talking about in the same way.
@amszmidt @theruran @50htz
Alfred M. Szmidt
@screwtape > later lisp-machines my understanding is that the microcoding of low level commands was actually done in lisp, as lisp, with lisp, even though the microcode itself was not lisp per se.
Not really.
screwlisp
@amszmidt
yduJ corrected me emphatically at one point that when she was doing it, it was done in lisp, with writing lisp. Though it's possible I have more than two misunderstandings.
@stman @theruran @50htz
★ STMAN ★ 🏳️🌈
@amszmidt @screwtape @theruran @50htz All this is very interesting to know for our culture, and for inspiration.
Actually the question this thread was trying to answer deals mainly with modeling. We were exploring a model-based (meaning model configured universal engine) universal engine and I asked how could these sub-models be, and why ? I wanted the justification for the splitting / creation of each sub-model.
screwlisp
@stman
Ah yeah, sorry for my off-topics
@amszmidt @theruran @50htz
Alfred M. Szmidt
@screwtape I don't know who that is, I do know how CADR, Lambda, K-machine, 36xx, Ivory, and Exploders work. One did not write microcode in Lisp in any shape or form.
Here is the implementation for the FALSE function (for CADR and Lambda/Exploder):
(MISC-INST-ENTRY FALSE)
XFALSE (POPJ-AFTER-NEXT (M-T) A-V-NIL)
(NO-OP)
This is not Lisp, it is microcode which just means "move A-V-NIL into M-T", and return.
★ STMAN ★ 🏳️🌈
Alfred,
Dunno if you had the patience to read my last posts, but we are revisiting the work done with hardcoded Lisp Machines in the 80', with the goal, this time, mainly of complexity reduction.
Our research group has already found 7 new concepts or design patterns for alternative digital system paradigms that enable by design their complexity reduction of several magnitude orders.
This time, we're trying to get rid of the
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.