5 ms·

I very much like this article. I especially like the comparison to TempleOS. I think Terry Davis blew computing wide open. That is what a personal computer can

I very much like this article. I especially like the comparison to TempleOS. I think Terry Davis blew computing wide open. That is what a personal computer can do, when it’s treated as an instrument instead of an appliance, and the user treated like the player of said instrument, instead of a child in a china shop. Imagine if you could plug your guitar into a computer, have that audio file, be a widget in a buffer, and then, pass it to other functions just like any other object, process it using, idk, map(car), and then sync it to a video using list functions, maybe with different implementations for the data structures under the hood. Or if you could make a game, modify it while you’re playing it, share that game as an image with a friend, and have them open it up, interact with it the same way they do with every other data structure from email to org mode.

ryukafalz 4y ago
Every time I see something like this, I'm also reminded of this demo of the Xerox Alto from a few years back. I'll link to the start of the most relevant piece: https://youtu.be/AnrlSqtpOkw?t=549 https://youtu.be/AnrlSqtpOkw?t=549 I'm always struck by the useful directions that desktop computers were going in the 70s, and how computing could be so much better than it is right now. But we've veered so off-course from all of this, away from composability and towards independent boxes ("applications") that can't talk to each other at all. Emacs is in some ways the best thing we have today in the other direction, but it pales in comparison to what's possible.
- vfclists 4y ago
  What Emacs users want is a version of Smalltalk with a superior keyboard driven experience, featuring which-key et al. > I'm always struck by the useful directions that desktop computers were going in the 70s, and how computing could be so much better than it is right now. But we've veered so off-course from all of this, away from composability and towards independent boxes ("applications") that can't talk to each other at all. This is the consequence of deliberate corporate agendas, but many people are still in denial of that. The worse thing that happened to this kind of computing was the policy of forcing everything through the web interface, which was deliberately crippled to let the industry majors to get a heads up in browser based software.
  - msla 4y ago
    > What Emacs users want is a version of Smalltalk with a superior keyboard driven experience, featuring which-key et al. No, more like Interlisp, which is kinda like Smalltalk in terms of being an active environment, only without having to deal with... you know... Smalltalk in order to use it. Plus, there's something to be said for a plan that's probably now mostly forgotten, to re-write Emacs to use Guile as its language and compile other languages to Guile so programmers could extend it in any of multiple languages and they'd all interoperate. The target language doesn't have to be Guile, of course; the modern conception would likely involve WASM and/or LLVM bitcode. https://www.emacswiki.org/emacs/GuileEmacs https://www.emacswiki.org/emacs/GuileEmacs
    - snamellit 4y ago
      Interestingly TJ Devries just did something like that for neovim. He made a transpiler from vim9script to Lua to allow compatibility. I found that around neovim there is a similar crafting vibe you find around Emacs.
      - vfclists 4y ago
        And they've got Fennel and some Fennel based frameworks to enable them to configure the Vim settings in a Lisp.
- abudabi123 4y ago
  > Emacs is in some ways the best thing we > have today in the other direction, but > it pales in comparison to what's > possible. Imagine joining Blender, sensors, actuators, voice assist, Ai, lisp, Emacs. If a project could attract the backers Blender is getting to explore and document design possibilites for Emacs that's a move in a better direction.
  - ykonstant 4y ago
    Yes, yes, yes and yes. I am glad to see this sentiment expressed here!
- agumonkey 4y ago
  I wonder if the sidefx houdini guys knew about smalltalk and if kay and his team ever saw houdini. The universal connecting data is very similar (and one of the most pleasurable thing ever)
- jll29 4y ago
  Very true. From one vantage point, things look like this: 1970s: The Visionaries - showing us how the future (of desktop computing) can look like. 2023: pre-1970s state stuff has become pervasive, runs faster and is cheaper, but not nearly at the level of sophistication that the 1970s had. (State of the art is still free clones of Unix, a 1960s OS.) So where is the course correction? And where are the 2023 ideas that as look far into the future from today as the J.C.R. Lickliders, D. Engelbarts, T. Nelson and A. Kays of the day? (Thankfully, in some areas, things look better - e.g. mobile phones, robotics, machine learning applications etc., so the point here is only valid about destop HCI progress.)
rtpg 4y ago
I find the fascination with TempleOS honestly weird, and I think people are honestly only fascinated by it because of the author. We have had smalltalk for the longest time. We have had so many attempts at "embed rich content into shell", we have had all of this! Hell, all of what you describe is basically available in the form of Mathmatica. And "share a program with a friend" is called a webpage with Javascript. We have the universal VM, just not in the shape that you would like. I don't want to be too dismissive of this stuff but I think at one point we gotta try using all these new fancy ideas that would be good for everyone, instead of constantly just imagining "what if they existed". Let's actualize a bit!
ykonstant 4y ago
>Imagine if you could plug your guitar into a computer, have that audio file, be a widget in a buffer, and then, pass it to other functions just like any other object, process it using, idk, map(car), and then sync it to a video using list functions, maybe with different implementations for the data structures under the hood. Amazing, thank you. This comment, and the replies to it, perfectly summarize what I am looking for in a complete LISP environment. True integration will treat not only text blocks/streams, but multimedia of any kind, seamlessly and as first-class citizens. The old LISP engineers understood that, but as another person said, the limitations of that time put significant barriers to that goal. We no longer have such limitations; it is time to revisit that model.

ryukafalz 4y ago

Every time I see something like this, I'm also reminded of this demo of the Xerox Alto from a few years back. I'll link to the start of the most relevant piece: https://youtu.be/AnrlSqtpOkw?t=549 https://youtu.be/AnrlSqtpOkw?t=549 I'm always struck by the useful directions that desktop computers were going in the 70s, and how computing could be so much better than it is right now. But we've veered so off-course from all of this, away from composability and towards independent boxes ("applications") that can't talk to each other at all. Emacs is in some ways the best thing we have today in the other direction, but it pales in comparison to what's possible.

vfclists 4y ago
What Emacs users want is a version of Smalltalk with a superior keyboard driven experience, featuring which-key et al. > I'm always struck by the useful directions that desktop computers were going in the 70s, and how computing could be so much better than it is right now. But we've veered so off-course from all of this, away from composability and towards independent boxes ("applications") that can't talk to each other at all. This is the consequence of deliberate corporate agendas, but many people are still in denial of that. The worse thing that happened to this kind of computing was the policy of forcing everything through the web interface, which was deliberately crippled to let the industry majors to get a heads up in browser based software.
- msla 4y ago
  > What Emacs users want is a version of Smalltalk with a superior keyboard driven experience, featuring which-key et al. No, more like Interlisp, which is kinda like Smalltalk in terms of being an active environment, only without having to deal with... you know... Smalltalk in order to use it. Plus, there's something to be said for a plan that's probably now mostly forgotten, to re-write Emacs to use Guile as its language and compile other languages to Guile so programmers could extend it in any of multiple languages and they'd all interoperate. The target language doesn't have to be Guile, of course; the modern conception would likely involve WASM and/or LLVM bitcode. https://www.emacswiki.org/emacs/GuileEmacs https://www.emacswiki.org/emacs/GuileEmacs
  - snamellit 4y ago
    Interestingly TJ Devries just did something like that for neovim. He made a transpiler from vim9script to Lua to allow compatibility. I found that around neovim there is a similar crafting vibe you find around Emacs.
    - vfclists 4y ago
      And they've got Fennel and some Fennel based frameworks to enable them to configure the Vim settings in a Lisp.
abudabi123 4y ago
> Emacs is in some ways the best thing we > have today in the other direction, but > it pales in comparison to what's > possible. Imagine joining Blender, sensors, actuators, voice assist, Ai, lisp, Emacs. If a project could attract the backers Blender is getting to explore and document design possibilites for Emacs that's a move in a better direction.
- ykonstant 4y ago
  Yes, yes, yes and yes. I am glad to see this sentiment expressed here!
agumonkey 4y ago
I wonder if the sidefx houdini guys knew about smalltalk and if kay and his team ever saw houdini. The universal connecting data is very similar (and one of the most pleasurable thing ever)
jll29 4y ago
Very true. From one vantage point, things look like this: 1970s: The Visionaries - showing us how the future (of desktop computing) can look like. 2023: pre-1970s state stuff has become pervasive, runs faster and is cheaper, but not nearly at the level of sophistication that the 1970s had. (State of the art is still free clones of Unix, a 1960s OS.) So where is the course correction? And where are the 2023 ideas that as look far into the future from today as the J.C.R. Lickliders, D. Engelbarts, T. Nelson and A. Kays of the day? (Thankfully, in some areas, things look better - e.g. mobile phones, robotics, machine learning applications etc., so the point here is only valid about destop HCI progress.)

rtpg 4y ago

I find the fascination with TempleOS honestly weird, and I think people are honestly only fascinated by it because of the author. We have had smalltalk for the longest time. We have had so many attempts at "embed rich content into shell", we have had all of this! Hell, all of what you describe is basically available in the form of Mathmatica. And "share a program with a friend" is called a webpage with Javascript. We have the universal VM, just not in the shape that you would like. I don't want to be too dismissive of this stuff but I think at one point we gotta try using all these new fancy ideas that would be good for everyone, instead of constantly just imagining "what if they existed". Let's actualize a bit!

ykonstant 4y ago

>Imagine if you could plug your guitar into a computer, have that audio file, be a widget in a buffer, and then, pass it to other functions just like any other object, process it using, idk, map(car), and then sync it to a video using list functions, maybe with different implementations for the data structures under the hood. Amazing, thank you. This comment, and the replies to it, perfectly summarize what I am looking for in a complete LISP environment. True integration will treat not only text blocks/streams, but multimedia of any kind, seamlessly and as first-class citizens. The old LISP engineers understood that, but as another person said, the limitations of that time put significant barriers to that goal. We no longer have such limitations; it is time to revisit that model.