Hi,
The Medley release of Xerox Interlisp-D has been open sourced under the MIT licence. Medley is/was an Interlisp/Common Lisp system that, like Smalltalk, ran on the Xerox D* machines and, also like Smalltalk, has a very interactive, image-based development style. I'm given to understand that it's very much like a Smalltalk environment using Lisp as it's language. Sorry that this is kinda off topic but I thought it would be interesting to Squeakers to know that a sort of sibling of the original Smalltalks had escaped from the propriety world like Squeak did :) More info here https://interlisp.org/ https://github.com/Interlisp |
Douglas, It's always good to see historically significant software preserved rather than bit rotting in darkness. It's probably not going to have nearly the impact Squeak did in the open source world since the Lisp world already has an embarrassment of good implementations (distinct implementations rather than dialects. See CCL, which is a direct descendant of MCL, for example) where Squeak basically had to introduce/resurrect Smalltalk in the open source world. With SLIME they've already kinda, sorta stolen some of the key concepts of the Lisp workstation UIs. Though this source drop may inspire some deeper copying on the UI front... we'll see. Thanks, Phil On Mon, Oct 12, 2020 at 12:16 PM Douglas Brebner <[hidden email]> wrote: Hi, |
On 12/10/2020 18:12, Phil B wrote: > Douglas, > > It's always good to see historically significant software preserved > rather than bit rotting in darkness. It's probably not going to have > nearly the impact Squeak did in the open source world since the Lisp > world already has an embarrassment of good implementations (distinct > implementations rather than dialects. See CCL, which is a direct > descendant of MCL, for example) where Squeak basically had to > introduce/resurrect Smalltalk in the open source world. With SLIME > they've already kinda, sorta stolen some of the key concepts of the > Lisp workstation UIs. Though this source drop may inspire some deeper > copying on the UI front... we'll see. Agreed. However, I think Medley is particularly interesting; it has some features and attributes that even Genera didn't have and isn't subject to the limitations of Emacs. It's also not an emulation like LambdaDelta, rather using a bytecode VM similar to Squeak. It may not make a big impact but I expect it will get a bunch of extremely dedicated fans. And, of course, it has a Smalltalk feel which is appealing :) |
On 2020-10-13 00:21,
Douglas Brebner wrote:
(My emphasis added) How I wish that the Smalltalk world had at least one good implementation (distinct implementation rather than dialect.) of ST80. Squeak is, of course, a moving target that leaves a trail of bit rotting software in darkness behind it. (sigh) Trygve --
The essence of object orientation is
that objects collaborate to achieve a
goal. |
Trygve Reenskaug wrote on Tue, 13 Oct 2020 10:37:38 +0200
> How I wish that the Smalltalk world had at least one good > implementation (distinct implementation rather than dialect.) > of ST80. Squeak is, of course, a moving target that leaves a > trail of bit rotting software in darkness behind it. (sigh) There have actually been two new virtual machines developed for Smalltalk-80 in the past year or so. Both are hobby projects, though with todays very fast computers even such can be very usable. I don't know the legality of using the old Xerox image, however. The image is museum quality: the exact bits it had in the early 1980s. But that is just because nobody is actually using it. For the Smalltalk-78 restoration project, for example, they started fixing and improving stuff and that is what Alan demoed. So I doubt that a Smalltalk-80 that a group actually used would change any less than Squeak does. There is always the option of using older Squeaks. I am typing this in Squeak 4.1 and for a project I wanted a very small image for I am using Squeak 2.2. It isn't always easy to find a virtual machine that will run on current operating systems and that can accept such old images. And obviously I have to do without more than two decades of bug fixes, but that is a price I am willing to pay. Pharo changes a lot more than Squeak does (the fork was so it could do so), but what about Cuis? I know its goal is simplicity and not historical stability but my impression is that you get a bit of that as well. -- Jecel https://github.com/dbanay/Smalltalk https://github.com/rochus-keller/Smalltalk |
Jecel, On Sun, Oct 18, 2020 at 9:19 PM Jecel Assumpcao Jr <[hidden email]> wrote: Pharo changes a lot more than Squeak does (the fork was so it could do Cuis goes for a small image and simplicity, but not backwards compatibility. It changes in backwards incompatible ways far more frequently and substantially than Squeak. For example, any Morphic code ported from Squeak would need to be significantly altered to run in Cuis.
Thanks, Phil |
In reply to this post by Jecel Assumpcao Jr
... and Xerox revivals start forking :
https://github.com/no-defun-allowed/Smalltalk/ https://github.com/michaelengel/crosstalk https://www.youtube.com/watch?v=JyaQavN9rVA cheers, Davide On 19/10/2020 03:19, Jecel Assumpcao Jr wrote: > Trygve Reenskaug wrote on Tue, 13 Oct 2020 10:37:38 +0200 >> How I wish that the Smalltalk world had at least one good >> implementation (distinct implementation rather than dialect.) >> of ST80. Squeak is, of course, a moving target that leaves a >> trail of bit rotting software in darkness behind it. (sigh) > There have actually been two new virtual machines developed for > Smalltalk-80 in the past year or so. Both are hobby projects, though > with todays very fast computers even such can be very usable. I don't > know the legality of using the old Xerox image, however. > > The image is museum quality: the exact bits it had in the early 1980s. > But that is just because nobody is actually using it. For the > Smalltalk-78 restoration project, for example, they started fixing and > improving stuff and that is what Alan demoed. So I doubt that a > Smalltalk-80 that a group actually used would change any less than > Squeak does. > > There is always the option of using older Squeaks. I am typing this in > Squeak 4.1 and for a project I wanted a very small image for I am using > Squeak 2.2. It isn't always easy to find a virtual machine that will run > on current operating systems and that can accept such old images. And > obviously I have to do without more than two decades of bug fixes, but > that is a price I am willing to pay. > > Pharo changes a lot more than Squeak does (the fork was so it could do > so), but what about Cuis? I know its goal is simplicity and not > historical stability but my impression is that you get a bit of that as > well. > > -- Jecel > > https://github.com/dbanay/Smalltalk > > https://github.com/rochus-keller/Smalltalk > Ing. Davide Grandi email : [hidden email] linkedin : http://linkedin.com/in/davidegrandi |
In reply to this post by Jecel Assumpcao Jr
On 19/10/2020 02:19, Jecel Assumpcao Jr wrote: > > There is always the option of using older Squeaks. I am typing this in > Squeak 4.1 and for a project I wanted a very small image for I am using > Squeak 2.2. It isn't always easy to find a virtual machine that will run > on current operating systems and that can accept such old images. And > obviously I have to do without more than two decades of bug fixes, but > that is a price I am willing to pay. I see you're using Celeste. Is it still being updated? And is it still under the Squeak-L licence like squeakmap says? I miss having tools that would let you "live" in a smalltalk image. |
Douglas Brebner wrote on Mon, 19 Oct 2020 11:30:09 +0100
> I see you're using Celeste. Is it still being updated? Earlier this year I looked into that as I was trying to updated my Celeste image from Squeak 4.1 32 bits to 5.3 64 bits in order to overcome the 4GB email limit. I found a few Celeste packages that were newer than what I had in 4.1. I also have a few modifications of my own, which overlap these newer packages quite a bit. Unfortunately reading my changeset files messed up things in 5.3 and I didn't investigate the problem further and instead reset my email archive, which has since grown to 358MB again. > And is it still > under the Squeak-L licence like squeakmap says? The following Celeste developers has signed a release to have their Squeak-L code available also under the MIT license: Adam Spitz, Daniel Vainsencher and Lex Spoon. On the other hand, the incomplete list I have of signatories doesn't include Bernhard Pieber nor Giovanni Giorgi, who is listed as the current administrator for Celeste on Squeak Source. Looking at old .changes files I can also see Celeste methods from Andreas Raab, Alejandro Magistrello, Vanessa Freudenberg, Dan Ingalls, Javier Diaz-Reinoso, John Maloney, Mike Rutenberg, Ned Konz, Bob Arning, Richard A. O'Keefe, Richard A. O'Keefe, Steve Elkins, Scott Wallace, Ted Kaehler and Yoshiki Ohshima. There were also methods by "dms" whose name I don't know. I think it is reasonable to consider that Celete is available under the MIT license. > I miss having tools that would let you "live" in a smalltalk image. Scamper started out a bit outdated and was quickly left behind as the web kept changing. Other things are a little more usable in today's world. A fully working SqueakNOS or equivalent would quickly show what is missing. -- Jecel |
Hi all! > Scamper started out a bit outdated and was quickly left behind as the web kept changing. Here is a more recent version of Scamper: Here is a MagicMouse, different approach on Web browsing, which uses Chromium and its PNG export plus UI event magic from Morphic to Chromium: Best, Marcel
|
In reply to this post by Jecel Assumpcao Jr
On 20/10/2020 23:15, Jecel Assumpcao Jr
wrote:
Douglas Brebner wrote on Mon, 19 Oct 2020 11:30:09 +0100I think it is reasonable to consider that Celete is available under the MIT license.
Thank you
I miss having tools that would let you "live" in a smalltalk image.Scamper started out a bit outdated and was quickly left behind as the web kept changing. Other things are a little more usable in today's world. A fully working SqueakNOS or equivalent would quickly show what is missing. Yes, web stuff, especially rendering, is a hideously complex,
massive task. Maybe an approach like the nyxt browser (common
lisp) which wraps an engine would be feasible but I'm not
confident. Special purpose clients should work, assuming they
don't assume the availiblility of a web renderer. |
Free forum by Nabble | Edit this page |