[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Fwd: Closer to Scheme in ROM

From: Eng. Stewart Jay Milberger
Subject: Fwd: Closer to Scheme in ROM
Date: Wed, 29 Jun 2022 13:32:50 +0000
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.11.0

-------- Forwarded Message --------
Subject: Re: Closer to Scheme in ROM
Date: Wed, 29 Jun 2022 13:22:30 +0000
From: Eng. Stewart Jay Milberger <>

Please forgive me not remembering to 'reply-all' yesterday, I woke up
in the clothes I fell asleep in, for reference.

Omitting the detail that I was in front of a computer, but the point has been made.  I am still in the same clothes, but now freshly washed.

(4 microseconds)

I thought would be good Knuth response, actually quite kind, as an answer "Do you have a couple of hours to talk about algorithms?"

That made me think how that could be abstracted with a variable ',units in Scmutils, when...Darth Internet...

Dante, speak not to me of Hell; Hugo, speak not to me of Quasimoto; luxury! compared to living and working on the 'net. "Breaky" and "non-performant-y", and "crashy-crashy" are now technical terms for why we keep Scheme around (and of course, the inferior Common Lisp and Dylan).

I don't know who replied to this thread personally (my tacit agreement with the world is "I don't know you, you don't know me" [1]), thank you for your alms to old Quasimoto here.  However, I have been acting in earnst--and realized we have hope for the ideas of Thomas and all Schemers and Lispers coexisting.  Viz S...O...S...the object system in MIT Scheme.

I wasn't even going to return here, but I found the C-style cast git commit in the microcode, and remembered, "I don't know why people who enjoy programming in Scheme would ever want to patch code in C."  Mr. Wu [I will sign an affidavit to attest that I don't know him, if requested.] made a fine contribution to the literature about how this could be accomplished, Scheme-86. Compromises, excuses, technical debt pushed to the future..leading to the world we have in general.

If you see me walking around Cambridge with a broom or a rake, it is to protect me from the feral dogs, as well as a protest to lawn mowers, leaf blowers, and Concorde Deux.  "Why not biplanes?", is the working slogan Dylan/Thomas Labs, unincorporated.

-- Stew J.

[1] Conversation with an old friend from Columbia about what to do with his aerospace degree, and the folly of trying to apply it to improved aviation safety in a concrete, rather than abstract way.

On 6/28/22 6:15 PM, Stewart J. Milberger wrote:
"Star* MMXXII: A New Hope"

Please forgive me not remembering to 'reply-all' yesterday, I woke up in the clothes I fell asleep in, for reference.

Is there a canonical VM instruction set for Dylan? I'd expect something simulataneously Smalltalk and Lisp flavored.

They are using code derived from that the Functional Objects company developed, now free software.  The original compiler is the DFMC, and had either native code via their HARP abstract RISC machine or portable C.  This seems similar to what MIT Scheme has, but I think MIT Scheme has a VM option now.

Peter Housel at Open Dylan has done a lot of work getting the DFMC connected to the LLVM toolchain, similar to Juila.

Could we hope for Open Dylan to revive this syntax and support it as an alternative to the infix one?

I'm doing a 6 month project called Dylan/Thomas Labs to look into it, as I would like Scheme to coexist with Dylan.  Hope is about all I have done so far.

Some papers on Dylan, Housel's 2020 paper has a diagram of the flow compiler (pg. 2).

reply via email to

[Prev in Thread] Current Thread [Next in Thread]