Jeff's works, survival.

I had written a while ago, asking about who had the rights to Jeff's code, like aha etc. To be able to use it as a drop in, but also as an extension to color forth, providing a GUI and target OS environment which could be translated to local GUI and OS, becoming a virtual target but, as a frame work API. This would give a user base to preserve Jeff's work, and it could be incorporated as a lite virtual environment in colorforth in ANSI Forth. But, I didn't hear back from anybody.
It's not really a matter of ANSI swallows colorforth and Jeff's work, but Chuck's and Jeff's work conquers ANSI from the inside out. :)

Anyway, what do people think, and who is in control of Jeff's code?

Thanks.

Wayne morellini <waynemorellini@gmail.com> writes:
> Anyway, what do people think, and who is in control of Jeff's code?

My guess is his widow is in control of the code, but it's probably more
interesting to adapt its ideas than the code itself. aha was among
other things extremely compact, so writing something like it wouldn't be
a huge task.

I'd also say that aha took a completely different direction from ANS,
and there is not much possibility of merging the two. The same holds
for Colorforth, which starts out by getting rid of colon definitions,
highly non-ANS ;). The most you could do is adapt some ideas here and
there from both aha and Colorforth. Jeff was a fertile thinker and
reading his stuff is always inspirational, but I think his code was so
specific to the environments he worked in, that trying to use it
directly wouldn't be fruitful. It's better to treat it as spiritual
guidance.

On 14/04/2022 18:33, Wayne morellini wrote:
> I had been suggesting before, that colorforth (machine forth) could be made a subset of ANSI, to keep it alive, along with a embedded version, and simple and complex additions (I forget exactly what then was, but basically extensions to handle extra stuff on the job I think). All subsets of ANSI, so ANSI lite color, embedded, as seperate word sets). It's in the ANN: colorforth cf2022 thread.

So ANSI should annex colorForth (for the latter's own good of course)?
Nations play that game too.

On Thursday, April 14, 2022 at 7:38:09 PM UTC+10, dxforth wrote:
> On 14/04/2022 18:33, Wayne morellini wrote:
> > I had been suggesting before, that colorforth (machine forth) could be made a subset of ANSI, to keep it alive, along with a embedded version, and simple and complex additions (I forget exactly what then was, but basically extensions to handle extra stuff on the job I think). All subsets of ANSI, so ANSI lite color, embedded, as seperate word sets). It's in the ANN: colorforth cf2022 thread.
> So ANSI should annex colorForth (for the latter's own good of course)?
> Nations play that game too.

Or, you should view it as inserting itself into ANSI. I'm aiming to talk about this in a reply to dxforth.

On Thursday, April 14, 2022 at 7:34:18 PM UTC+10, Paul Rubin wrote:
> Wayne morellini <waynemo...@gmail.com> writes:
> > Anyway, what do people think, and who is in control of Jeff's code?
> My guess is his widow is in control of the code, but it's probably more
> interesting to adapt its ideas than the code itself. aha was among
> other things extremely compact, so writing something like it wouldn't be
> a huge task.
>
> I'd also say that aha took a completely different direction from ANS,
> and there is not much possibility of merging the two. The same holds
> for Colorforth, which starts out by getting rid of colon definitions,
> highly non-ANS ;). The most you could do is adapt some ideas here and
> there from both aha and Colorforth. Jeff was a fertile thinker and
> reading his stuff is always inspirational, but I think his code was so
> specific to the environments he worked in, that trying to use it
> directly wouldn't be fruitful. It's better to treat it as spiritual
> guidance.

I'm really saying, as a secondary environment inside ANSI. You code fur either environment, or can swap between them. Both get standard maintenance, and both can share some underlying infrastructure, as needed. It's really a way of having ANSI Colorforth Engine, without having a seperate ANSI standards to too many committees. Frankly, I agree that having familiar C like file handling etc for C types, is nice, but they should have developed their own parallel handling routines. People can hold onto the past, or come into the future. But here, we can partition the functionality in forth, to basically have a number of seperate systems in the same language, which forth allows. But not going over board. We need low resource embedded, and graphical GUI personal interface like handling. We need to do what C doesn't.

Now, from the above we see we can parallel include colorforth and aha, sneakily, to the benefit of both.

Now, you remember they designed machine forth to also run in a native code 386 code version, this can be extended fur many processors, or use tokens, either way, the code itself can run, but the routines to draw stuff will have to be adjusted to the local target environment, but that's normal. The compiler will have to be made to work with the local is and hardware in any case. Just like normal programs, the back end has to be customised to the environment, but with this, just the compiler has to be customised to the different target, and depending on the environment different code sets and backends are used. In reality, most significant things GUI will be using ARM, x86 and maybe Riscv in future. The strategy would be to load portable code for the first two processors at first, and the back end routines to use the environment through the OS (windows, Mac OS, Linux etc) at first. If there is any other environment, you specify the behaviour of the back end, and whoever comes along, can implement the native routines themselves. It gets around most 9f the custom misc MCU hardware stuff.

However, there has been too much wasting of time on the exotic dude of things 8n misc. Very simply, an 8 bit, 16 bit and 32 bit misc cells with full address space, and configuration with all the additional misc stuff, and video sound, memory and IO versions, is a seller, as single processor cells. As for parallel, a more advanced version with low overhead communications and ability to pass information to 8, neighbours, or anywhere on volume, rows, diagonal and globally. This can be done with a seperate bus for each, co-joined multi-bus memory, and a universal storage and swap memory. Some very simplified DMA, will keep the processor idle a lot more unless it's doing l work. The co-joined memory allows junks of processing to be done and swapped by sending a pointer. A lot more relatable to ordinary programmers.

People say how simple forth is, it simple to them, in the same way that Chuck might think how simple the GA144 might be. They don't see that they have mastery and/or know it, therefore it appears simpler to them. Fir people who don't know it, and can't master it easily, it's not. This is an industry where just being different might scare people away. So, you notice how I say, yeah, C file handling style in ANSI Forth. And yeah, we should have a parallel better forth style as well, it's to get over this delusional timidity resistance problem. You can't less a horse to drink, of whatever the saying was. But, you can say, here you can drink out of your bucket if you want, but there is a nice trough over here when your ready. Those people will pass on eventually, and the newer people trained on the new forth version, will take over. I notice there is not much strategic thinking like this here. As much as I might not appreciate the technical abilities of certain tech mogul, but now I look at it, they probably all were good at strategy. I could make IBM, the richest computer company in the world again, within 10 years, by strategy with better technology and delivery. Often when I read of failures, it's not just because they can't design or code things, and don't know what is good code or design, it's "why did they do that" bad strategy. I mean, I have been reviewing the history of the home computers and earlier console wars, in the 1980's going into the early 1990's, because I want to introduce a new retro home computer system, and it's just interesting. The Commodore Pet was a good example, that the person in charge bought on Chuck Peddle, and other strategists, who had been waiting to kick off home computing for many years. They convinced the new owner. A great business strategist, and did the Pets, then did the Vic 20 on an old Vic chip (something about the Vic 2 not been ready yet). They then brought our there intended Vic 2 computer, and it went ahead. Chuck left, and Jack left Commodore, and they pushed out the plus4 and commodore 128. Jack was very tight, and so was the chairman of the board, who Jack needed for money, which is an issue. At the time, the business market had been cpm, and the PC was very expensive. The previous commodore 500 business machine would have stood a chance against cpm crowd, but times suddenly changed. The plus 4 strategy was tight fisted and with Jack and Chuck gone, nobody was there to tell them, this isn't going fly anymore. With Chuck etc gone, I think, there would ha e been nobody there to point out the Commodore 128 was ham fisted, and they needed to out the money into developing a vic 3 chip and new computer, even if it was a 80 text column VGA extension of the Vic 2 chip with 16 sprites and 2-4 times the graphics data through put, and to upgrade the foundry to do 4-8 MHz in future, and try a dram or SRAM line. I feel they got stuck between past cost savings measures and nobody to tell them what to do after those. If Jack hadn't lost the Amiga negotiations, Commodore could have been toast a lot sooner, even within years the writing could be on the wall. The Atari 7800 architecture, was actually external, and the best. If they had added a novix to it, or put in proper effort, instead of delay it fur years, they could have had dominion in the higher and lower end at Atari, with Amiga, and 7800 architecture on a low end 65xx. As it was. Atari only extended its own life, by getting the people designing the Sinclair super spectrum on board, which was intentionally, like a stripped down Amiga competitor, where they deliberately designed it without duplicate methods for doing the same thing. This is where they latter hit the Jaguar architecture from, which was hobbled with an optional 68xxx chip.

See, strategy gets you out of these situations onto something better. Here, a better misc strategy is needed. Technically, me and Jeff, might be a bit similar end objectives at least, it might be a conservatism in action and energy thing from material arts, leading to efficiency of force. Bruce Lee also used to teach in that, and I don't know if my teacher taught on that there (he has common descent of teaching from IP Man, I think, who apparently was the source of Bruce's early teachings on conservative of energy). But somehow, I never saw it as picking the cut leaves our of the hedges, but something more than that, to her the leaves out, and design it so you didn't need to get the leaves out, as there was no leaves to get out. One approach is like a precise version of brute force, as far as design goes, the other, is conquering the design, to set a new architecture then paradigm. It is this way with strategy, to find victory in both the small and large way, done right.

In article <87h76wdn54.fsf@nightsong.com>,
Paul Rubin <no.email@nospam.invalid> wrote:
>Wayne morellini <waynemorellini@gmail.com> writes:
>> Anyway, what do people think, and who is in control of Jeff's code?
>
>My guess is his widow is in control of the code, but it's probably more
>interesting to adapt its ideas than the code itself. aha was among
>other things extremely compact, so writing something like it wouldn't be
>a huge task.
>
>I'd also say that aha took a completely different direction from ANS,
>and there is not much possibility of merging the two. The same holds
>for Colorforth, which starts out by getting rid of colon definitions,
>highly non-ANS ;). The most you could do is adapt some ideas here and
>there from both aha and Colorforth. Jeff was a fertile thinker and
>reading his stuff is always inspirational, but I think his code was so
>specific to the environments he worked in, that trying to use it
>directly wouldn't be fruitful. It's better to treat it as spiritual
>guidance.

Anyway, this is futile talk, unless aha is somewhere available.
I applaud that the widow would do that.

Groetjes Albert
--
"in our communism country Viet Nam, people are forced to be
alive and in the western country like US, people are free to
die from Covid 19 lol" duc ha
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

On Friday, April 15, 2022 at 3:01:10 AM UTC-5, none albert wrote:
> Anyway, this is futile talk, unless aha is somewhere available.

But it did induce some futile thoughts of my own but I will spare everyone.
--
me

On Saturday, April 16, 2022 at 2:46:32 AM UTC+10, S Jack wrote:
> On Friday, April 15, 2022 at 3:01:10 AM UTC-5, none albert wrote:
> > Anyway, this is futile talk, unless aha is somewhere available.
> But it did induce some futile thoughts of my own but I will spare everyone.
> --
> me
Oh, go on?

Am 14/04/2022 um 11:34 schrieb Paul Rubin:
> Colorforth, which starts out by getting rid of colon definitions
Hi Paul,

Just a on a technical note : colorForth still has colon definitions -
but they are defined by the colour of the token (red), not a ':'.
They can be displayed with a ':' if you select legacy/colour-blind mode,
so it looks the same as ANS Forth. Does that count?

Cheers,
Howerd

On Thursday, April 14, 2022 at 3:39:39 PM UTC+2, Wayne morellini wrote:
> This is an industry where just being different might scare people away.
> So, you notice how I say, yeah, C file handling style in ANSI Forth.
> And yeah, we should have a parallel better forth style as well, it's to get
> over this delusional timidity resistance problem. You can't less a horse
> to drink, of whatever the saying was.

Well, such has been developed and has been in use for several decades.
Although it was more the result of an fortunate accident than actual design.

4tH was initially meant to be a scripting language - something that was
essentially a PART of another application. Hence, not much file handling was
required. As a matter of fact, it was just a simple redirect from console I/O.

However, as 4tH developed it required a proper I/O system WITHOUT
breaking too much code. So:

1) TYPE was used as block-style output, since you already had an address
and a count;
2) REFILL was used for line based input (filling the TIB);
3) ACCEPT was used for block-style input, since you already had an address
and a count.

Now NONE of these have a file descriptor. So, 4tH simply expanded the
"redirection" idea by creating streams and channels (CIN and COUT). Once
you connect an open stream with a channel - that's where the output goes.

So { S" text.txt" INPUT OPEN } doesn't open a channel, but a stream. By
issuing { <fd> USE } the stream is connected to the proper channel. You don't
have to issue which channel, because you implicitly defined that when using
OPEN.

You can temporary redirect back to e.g. the screen by issuing:

COUT >R STDOUT USE ." Hello world!" CR R> USE

IMHO it has proven to be a very simple and very effective mechanism. E.g.
I wrote a simple DBMS around it. It's dead easy to define FILE wordset in it -
although the inverse is virtually impossible. There is NO distinction between
text files or binary files. Treat it as a binary file, you get a binary file. Treat it
as a text file and you get a text file.

Of course you could get I/O errors - and they are NOT returned as values.
OPEN is the only I/O related word that does this. All others just THROW an
exception.

Yeah, one could get a "disk full" error easily when using floppies - but that is not
the case anymore. So in my experience I/O exceptions are very, very rare IRL.

{ <fd> CLOSE } will close a file - but NOT return a value. It will reset the stream
to its default stream (being STDIN or STDOUT). If the stream is connected to
a channel, then that channel will "inherit" the reconnected stream.
Again, it's rare to see an I/O exception in this stage of file access.

And again, I'm not promoting 4tH's I/O model here. I just wanna say that alternative
I/O methods have been developed and in use for many, many years.

Hans Bezemer

Howerd Oakford <howerd@inventio.co.uk> writes:
> colorForth still has colon definitions -... They can be displayed
> with a ':' if you select legacy/colour-blind mode ... Does that count?

Ah ok, I guess so, though I guess you couldn't use cf's screen editor.
Also, macros (black words iirc) work completely differently from
traditional Forth. Can legacy mode help with that? There has to be
strict phase separation between compile time and run time, since the
target cpu has no text interpreter.

On Friday, April 15, 2022 at 1:53:22 PM UTC-5, Wayne morellini wrote:

> Oh, go on?

Ok, but just one.
Banking 101 or 'the game of musical chairs':
All enterprise are doomed to failure. Make the money while you can
(investors playing the long shot), get out before the music stops
(investors wising up) and leave the debt to someone else, preferably
the tax payer.
--
me

On Sun, 3 Apr 2022 18:45:41 -0700 (PDT)
Kip Ingram <kip.ingram@gmail.com> wrote:

> > >Anti Ansi Forth:
> > >
> > ><http://www.UltraTechnology.com/antiansi.htm>
> > >
> > >Leo Wong
> > >he...@albany.net
> > >http://www.albany.net/~hello/

Jeff details some of the work on ANS Forth for P21 here:

http://www.ultratechnology.com/p21fchp3.html

It is a pity that Jeff's work is not dated, as then we might understand
how his opinions developed a little better.

The instruction set for the P21 is 1/3 way down this page:

http://www.ultratechnology.com/p21fchp9.html
Jan Coombs
--

On Sunday, April 17, 2022 at 4:34:09 AM UTC+10, S Jack wrote:
> On Friday, April 15, 2022 at 1:53:22 PM UTC-5, Wayne morellini wrote:
>
> > Oh, go on?
>
> Ok, but just one.
> Banking 101 or 'the game of musical chairs':
> All enterprise are doomed to failure. Make the money while you can
> (investors playing the long shot), get out before the music stops
> (investors wising up) and leave the debt to someone else, preferably
> the tax payer.
> --
> me

I'm not interested in such games, but as time goes on and things are not easy, I'm looking at flipping off done things, rather than target it all under one roof eventually. But, there are very long term enterprises that survive. Failure is often circumstances, betrayal or competition, on incompetence. So, structure, talent, talent replenishment mechanism, buffering and good complete mechanisms, are way to sustain an enterprise. I'm no longer at the level I was, but a chairman works on the strategy for these things with the CEO, who then carries things out with the team. Steve Jobs had very little, and if he can make it, why not. We are so negative around here, we dint make it, every improvement is virtually there to be sent to the wolves. The lambs make the flock, wolves destroy things. I see ANSI, even GA, as an opportunity. When I say opportunity, I also mean to be fibrin such a way as to bring about the opportunity. I saw Forth OS, as an opportunity. But, the world is more than us, so, I advocate the JavaScript interface as a minimal way to cheaply reach compatibility with hardware and systems out there, and reach people. It's all a strategy to be affective. Problem is, people are very problemed seeing strategy. I've had a lot of success reaching out to companies with design and strategy since maybe 15, but truthfully, some biggest companies (or one on particular) are worse at implementing those designs and strategies. We face the issue here in the past, with people whose strategy is to take down others. In such a environment, genuine strategy problems are not given the attention they deserve. One such thing, is that the F21, was the last minimal instruction set computer device in forth, where as everything since has been even more minimum. We would have been better off crowd sourcing development of the F21 and arrays there of, back after Jeff passed, and develop 32 bit and maybe 64 bit versions.. But, also deep stacks, to nest routines, and a pointer passing to allow data to be dumped into memory, but be treated as on the stack when the pointer in encountered. This means one or more extra bits on each stack value. Anyway, I just came up with that just then, and haven't worked out the ISA mechanism it would work by. But, I have come up with simplified DMA, caching, and direct messaging systems requiring minimal circuitry. These aren't the multiple bus/memory communications pools I've mentioned before, but they are also minimal circuitry and timing conflict. The simplest thing about reducing conflicts and wait states, is to program to only use it when you need to. I regard poor code to be the problem of the developer, if their app tanks, in my systems, it is designed to be isolated enough from other apps and the system, to maybe hardly ever effect them much, and hardly effect them at all. So, the bad app appears to be the one causing the problem, and the ones who aren't function correctly. But, saying that, architecture, style guides, and training to minimise the bad apps problems, and why so advocate the worst programmers be restricted to writing web simple pages, to get them out of disrupting app and systems programming. It's a management strategy. This means on percents more transistors per present misc cores, could have big benefits in their areas. If we look at the bigger package from years ago, we are talking double percentage points, plus whatever memory, and some traces lines per direction on each row/column and maybe diagonal of chips. This might seem a waste per die, but you can only push so much in parallel, and these things use power doing useful work, and if you want higher performance the processors have to run faster, and having less density of chips per die allows thermal management to run them fast. The minimum energy on the full package are still likely to be about the lowest energy in low energy mode, but per unit of performance lower than the current misc design, just the bottom energy will be lower of the current misc design, due to its lower complexity. So applications, like waiting for odd signals to arrive over time, where little work is done, would suite present misc more. I think those applications, are very long term embedded like, waiting for port events etc. I would like to see the chip redesigned for thermal management. Square chips are a bad choice for this. On passive hear sink technology it's worse. I once proposed a Mac mini where the chips where arranged in an array as heat sink fins. So, each column and row, is affectively a long die with air passing over. Thousands of arm dues could fit in a grill like structure this way. My suggestion for the Mac Pro, seemed to gain traction, but within a year or so, they changed to the present scheme, which is really unfortunate. I was suggesting a way where mini Mac blades (not the insane 1kw cooker version mentioned before) could be stacked along along the desk, up, or by their hundreds as an array, with no cage. The present Mac Pro can't reach that performance, plus mine was aimed at the rendering and server markets as well as pro and mini. The modules could carry what ever card functionality you wanted. You can do with one mini (or seperate modules) to deal with human input, and the rest can be cpu card modules, etc. You practically can have hundreds of the biggest intel or arm chips running in an array. Now, we are approaching a time where this would only be needed in the most extreme cases. GPU's already are hitting towards high end performance for most consumer, and professional, ashes numbers/use cases. Still a 3-4 years sales advantage, and maybe even a decade life. So, still financially very valuable.

On Saturday, April 16, 2022 at 10:25:55 PM UTC-5, Wayne morellini wrote:
> On Sunday, April 17, 2022 at 4:34:09 AM UTC+10, S Jack wrote:
> > On Friday, April 15, 2022 at 1:53:22 PM UTC-5, Wayne morellini wrote:
> >
The game goes its way regardless of anyone's interest. Want to be a
player go with the game; otherwise, drop out or be left out.

Failure is often, period. The reasons matter not; they will be sundry
and varied like the weather.

Strategy doesn't win chess games. Your opponent won't be playing your
game. What wins is making the better moves for the situations at
hand.

Steve Job. Although the greatest superpower is luck, next is the gift
of gab. Skill is low on the list if there at all.

JavaScript as an interface, why? Forth has no interface problems.
Hardware interface is Forth's strong point and has nothing to do with
JS. All the web stuff, HTML XML SVG JSON etc., is just plain text.
Even Basic can handle that.

ANS is doing good just to meet its purpose, aiding a couple of vendors
and a handful of contractor to cope with maintaining software on
variety of hardware. If it does that, well and good. Don't expect
people to drop what their working with and embrace Forth just because
it has a standard.
--
me

On Sunday, April 17, 2022 at 4:29:46 PM UTC+10, S Jack wrote:
> On Saturday, April 16, 2022 at 10:25:55 PM UTC-5, Wayne morellini wrote:
> > On Sunday, April 17, 2022 at 4:34:09 AM UTC+10, S Jack wrote:
> > > On Friday, April 15, 2022 at 1:53:22 PM UTC-5, Wayne morellini wrote:
> > >
> The game goes its way regardless of anyone's interest. Want to be a
> player go with the game; otherwise, drop out or be left out.

That's not true. The players set the fame as much as they can. Only bit players think differently. Good grief, I just saw a Wozinak interview where he says he went to design a universal remote control, and got the company designing fur Apple, who takes new customers, to do sample remotes. Steve's cones by in his bike, the guy at the company shows him what Wozinak was doing, Jobs through the remote against the wall in this story, and told the guy that everything that company does, belongs to Apple, and to send all the sample remote designs Wozinak ordered in a box to his office. Bit the way to would act or condone. It's Wozinak's exclusive business with a third party, yet this is what can happen in the game. The chess board rearranged..

>
> Failure is often, period. The reasons matter not; they will be sundry
> and varied like the weather.

I think I outlined thT. And how to hedge against failure. Again, long term successful people usually don't, and as far as I know, don't, think like that.

>
> Strategy doesn't win chess games. Your opponent won't be playing your
> game. What wins is making the better moves for the situations at
> hand.

I don't think people going through thousands or tens of thousands of of chess moves in their head to pick the best strategy, is not strategy. Again. Really successful companies tend to set the hand strategically, to win repeatedly. How many years has IBM been going?
>
> Steve Job. Although the greatest superpower is luck, next is the gift
> of gab. Skill is low on the list if there at all.

Steve Jobs walks into an elevator, and sees what you are wearing, and says your fired. He was all about strategy design etc, I don't think he would be impressed about by luck.

>
> JavaScript as an interface, why? Forth has no interface problems.
> Hardware interface is Forth's strong point and has nothing to do with
> JS. All the web stuff, HTML XML SVG JSON etc., is just plain text.
> Even Basic can handle that.

You just called SVG plain text

Forth has a lot of interface issues where you have to design the interface, rather than use an API. Interfaces is not just some but banging port and text display, that's next to nothing. Interface is about an API and driver set that provides portable coding ability without having to custom make drivers and APIs to address the user, the system, other apps and systems, and the hardware.

>
> ANS is doing good just to meet its purpose, aiding a couple of vendors
> and a handful of contractor to cope with maintaining software on
> variety of hardware. If it does that, well and good. Don't expect
> people to drop what their working with and embrace Forth just because
> it has a standard.
..
Who said that? You would need a strategy for a good version that would attract people.

Thanks.

On Sunday, April 17, 2022 at 7:44:27 AM UTC-5, Wayne morellini wrote:
> You just called SVG plain text
>
Yes, I did.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<svg
xmlns:svg="http://www.w3.org/2000/svg"
xmlns="http://www.w3.org/2000/svg"
width="100%"
height="100%"
viewBox="0 0 640 480"
style="background-color:dimgray"
id="svg2"
version="1.1"
>
<script type="application/ecmascript"><![CDATA[
function clickOpacity(evt) {
var path = evt.target
var opacityValue = path.getAttribute("opacity");
if ( opacityValue == 0.0 )
path.setAttribute("opacity",1.0);
else
path.setAttribute("opacity",0.0);
} ]]></script>
<g id="g001" transform="translate(180,180)" >
<path
id="path001"
style="fill:silver;fill-rule:evenodd;stroke:none"
d="m 060, 040
100, -020
000, 100
-100, -020
z"
onclick="clickOpacity(evt);"
/>
<path
id="path002"
style="fill:red;fill-rule:evenodd;stroke:none"
d="m 000, 020
60, 020
000, 060
-60, 020
z"
onclick="clickOpacity(evt);"
/>

....

<path
id="path006"
style="fill:magenta;fill-rule:evenodd;stroke:none"
opacity="0.0"
d="m 000, 020
120, -020
000, 140
-120, -020
z"
onclick="clickOpacity(evt);"
/>
</g>
</svg>

--
me

On Monday, April 18, 2022 at 12:02:37 AM UTC+10, S Jack wrote:
> On Sunday, April 17, 2022 at 7:44:27 AM UTC-5, Wayne morellini wrote:
> > You just called SVG plain text
> >
> Yes, I did.
>
> <?xml version="1.0" encoding="UTF-8" standalone="no"?>
> <svg
> xmlns:svg="http://www.w3.org/2000/svg"
> xmlns="http://www.w3.org/2000/svg"
> width="100%"
> height="100%"
> viewBox="0 0 640 480"
> style="background-color:dimgray"
> id="svg2"
> version="1.1"
> >
> <script type="application/ecmascript"><![CDATA[
> function clickOpacity(evt) {
> var path = evt.target
> var opacityValue = path.getAttribute("opacity");
> if ( opacityValue == 0.0 )
> path.setAttribute("opacity",1.0);
> else
> path.setAttribute("opacity",0.0);
> }
> ]]></script>
> <g id="g001" transform="translate(180,180)" >
> <path
> id="path001"
> style="fill:silver;fill-rule:evenodd;stroke:none"
> d="m 060, 040
> 100, -020
> 000, 100
> -100, -020
> z"
> onclick="clickOpacity(evt);"
> />
> <path
> id="path002"
> style="fill:red;fill-rule:evenodd;stroke:none"
> d="m 000, 020
> 60, 020
> 000, 060
> -60, 020
> z"
> onclick="clickOpacity(evt);"
> />
>
> ...
>
> <path
> id="path006"
> style="fill:magenta;fill-rule:evenodd;stroke:none"
> opacity="0.0"
> d="m 000, 020
> 120, -020
> 000, 140
> -120, -020
> z"
> onclick="clickOpacity(evt);"
> />
> </g>
> </svg>
>
> --
> me

https://www.w3.org/Graphics/SVG/

Don't push it. It's a vector graphics language. It's origins, from the CSIRO graphics research lab patent, also used for flash, which came from a job application I submitted to the lab, with the idea attached, which so had come up with to transmit buttons icons and graphics over very low speed connections for my operating system project, and then to speed up the internet (the OS was meant to replace java and user side web scripting). Which, I'm not too happy about. No job interview, no acknowledgment, only read in the paper, hey they are applying for a patent on it several months later. .i get around a lot.

On Monday, April 18, 2022 at 1:33:28 AM UTC+10, Wayne morellini wrote:
> On Monday, April 18, 2022 at 12:02:37 AM UTC+10, S Jack wrote:
> > On Sunday, April 17, 2022 at 7:44:27 AM UTC-5, Wayne morellini wrote:
> > > You just called SVG plain text
> > >
> > Yes, I did.
> >
> > <?xml version="1.0" encoding="UTF-8" standalone="no"?>
> > <svg
> > xmlns:svg="http://www.w3.org/2000/svg"
> > xmlns="http://www.w3.org/2000/svg"
> > width="100%"
> > height="100%"
> > viewBox="0 0 640 480"
> > style="background-color:dimgray"
> > id="svg2"
> > version="1.1"
> > >
> > <script type="application/ecmascript"><![CDATA[
> > function clickOpacity(evt) {
> > var path = evt.target
> > var opacityValue = path.getAttribute("opacity");
> > if ( opacityValue == 0.0 )
> > path.setAttribute("opacity",1.0);
> > else
> > path.setAttribute("opacity",0.0);
> > }
> > ]]></script>
> > <g id="g001" transform="translate(180,180)" >
> > <path
> > id="path001"
> > style="fill:silver;fill-rule:evenodd;stroke:none"
> > d="m 060, 040
> > 100, -020
> > 000, 100
> > -100, -020
> > z"
> > onclick="clickOpacity(evt);"
> > />
> > <path
> > id="path002"
> > style="fill:red;fill-rule:evenodd;stroke:none"
> > d="m 000, 020
> > 60, 020
> > 000, 060
> > -60, 020
> > z"
> > onclick="clickOpacity(evt);"
> > />
> >
> > ...
> >
> > <path
> > id="path006"
> > style="fill:magenta;fill-rule:evenodd;stroke:none"
> > opacity="0.0"
> > d="m 000, 020
> > 120, -020
> > 000, 140
> > -120, -020
> > z"
> > onclick="clickOpacity(evt);"
> > />
> > </g>
> > </svg>
> >
> > --
> > me
> https://www.w3.org/Graphics/SVG/
>
> Don't push it. It's a vector graphics language. It's origins, from the CSIRO graphics research lab patent, also used for flash, which came from a job application I submitted to the lab, with the idea attached, which so had come up with to transmit buttons icons and graphics over very low speed connections for my operating system project, and then to speed up the internet (the OS was meant to replace java and user side web scripting). Which, I'm not too happy about. No job interview, no acknowledgment, only read in the paper, hey they are applying for a patent on it several months later. .i get around a lot.

I'm, not saying I was original in that, as I think there might have been a bulletin board software package which used a vector graphics drawing, but still, worth a patent, worth a mention.

On Sunday, April 17, 2022 at 5:07:54 AM UTC+10, Jan Coombs wrote:
> On Sun, 3 Apr 2022 18:45:41 -0700 (PDT)
> Kip Ingram <kip.i...@gmail.com> wrote:
>
> > > >Anti Ansi Forth:
> > > >
> > > ><http://www.UltraTechnology.com/antiansi.htm>
> > > >
> > > >Leo Wong
> > > >he...@albany.net
> > > >http://www.albany.net/~hello/
>
> Jeff details some of the work on ANS Forth for P21 here:
>
> http://www.ultratechnology.com/p21fchp3.html
>
> It is a pity that Jeff's work is not dated, as then we might understand
> how his opinions developed a little better.
>
> The instruction set for the P21 is 1/3 way down this page:
>
> http://www.ultratechnology.com/p21fchp9.html
>
> Jan Coombs
> --

Hi Jan.

Did Jeff ever reveal the instruction set for the F21? It would be good to examine his thinking on refining the instruction set, as he did a lot of simulation testing on it.

Actually, I have a though for an official open source Forth processor thread

On Monday, April 18, 2022 at 1:07:02 AM UTC-4, Wayne morellini wrote:
> Actually, I have a though[t] for an official open source Forth processor thread

Before you do that, I suggest that you:
a) install Icarus Verilog and learn to use it
b) define a version0 instruction set, CPU registers, and interface signals
c) learn Verilog by achieving successful compilations of all your hardware definition modules
d) perform adequate Verilog simulations of all modules, including the top-level CPU
e) publish your Verilog source and test vector results

This will take some time.

- Myron

On Sun, 17 Apr 2022 22:07:00 -0700 (PDT)
Wayne morellini <waynemorellini@gmail.com> wrote:

> Did Jeff ever reveal the instruction set for the F21? It would be good to examine his thinking on refining the instruction set, as he did a lot of simulation testing on it.

F21 Microprocessor Preliminary specifications 9/98

Code | Name | Description | Forth (with var named A)
-----+------+------------------------------+--------------------------
00 | else | Unconditional jump | ELSE
01 | T0 | Jump is T0-19==0 | DUP IF
02 | call | Push PC+1 to R, jump | :
03 | C0 | Jump if T20 is false | CARRY? IF
06 | RET | Pop PC from R (return) | ;
08 | @R+ | Fetch from addr in R, inc R | R@ @ R> 1+ >R
09 | @A+ | Fetch from addr in A, inc A | A @ @ 1 A +!
0A | # | Fetch from PC+1, incr PC | LIT
0B | @A | Fetch from address in A | A @ @
0C | !R+ | Store to addr in R, inc R | R@ ! R> 1+ >R
0D | !A+ | Store to addr in A, inc A | A @ ! 1 A +!
0F | !A | Store to addr in A | A @ !
10 | com | Complement T | -1 XOR
11 | 2* | Left shift T, 0 to T0 | 2*
12 | 2/ | Right shift T, T20 to T19 | 2/
13 | +* | Add S to T if T0 is true | DUP 1 AND IF OVER + THEN
14 | -or | Exclusive-or S to T | XOR
15 | and | And S to T | AND
17 | + | Add S to T | +
18 | pop | Pop R, push to T | R>
19 | A | Push A to T | A @
1A | dup | Push T to T | DUP
1B | over | Push S to T | OVER
1C | push | Pop T, push to R | >R
1D | A! | Pop T to A | A !
1E | nop | Delay 2ns | NOP
1F | drop | Pop T | DROP

abridged from:
f21data__F21_PreliminarySpecifications_1998-09.pdf

It would be good to be able to put this family of chips in date order,
and note which ones made it to fully working silicon.

The Mup21 was available in a PLCC44 package to buy in 1995. The video
processor on the F21 is likely derived from the MuP21.

> Actually, I have a though for an official open source Forth processor thread

Good, even for commercial gain the open source route is probably most effective.

Jan Coombs
--

On Monday, April 18, 2022 at 9:41:27 PM UTC+10, Myron Plichota wrote:
> On Monday, April 18, 2022 at 1:07:02 AM UTC-4, Wayne morellini wrote:
> > Actually, I have a though[t] for an official open source Forth processor thread
>
> Before you do that, I suggest that you:
> a) install Icarus Verilog and learn to use it
> b) define a version0 instruction set, CPU registers, and interface signals
> c) learn Verilog by achieving successful compilations of all your hardware definition modules
> d) perform adequate Verilog simulations of all modules, including the top-level CPU
> e) publish your Verilog source and test vector results
>
> This will take some time.
>
> - Myron
[Sarcastically] Oh yes. You think it's not about community, but you think it's all about me.

Why don't you go and do it, and contribute it back to the community.

Myron doesn't realise what I'm suggesting is not a design I would make, it is community. But it never stops the never help, from bathing in and not contributing positively. My own designs are a lot different and commercial, so a commercial agreement would be needed for those to be used.

On Monday, April 18, 2022 at 8:00:16 AM UTC-4, Jan Coombs wrote:
> F21 Microprocessor Preliminary specifications 9/98
>
> Code | Name | Description | Forth (with var named A)
> -----+------+------------------------------+--------------------------
> 00 | else | Unconditional jump | ELSE
> 01 | T0 | Jump is T0-19==0 | DUP IF
> 02 | call | Push PC+1 to R, jump | :
> 03 | C0 | Jump if T20 is false | CARRY? IF
> 06 | RET | Pop PC from R (return) | ;
> 08 | @R+ | Fetch from addr in R, inc R | R@ @ R> 1+ >R
> 09 | @A+ | Fetch from addr in A, inc A | A @ @ 1 A +!
> 0A | # | Fetch from PC+1, incr PC | LIT
> 0B | @A | Fetch from address in A | A @ @
> 0C | !R+ | Store to addr in R, inc R | R@ ! R> 1+ >R
> 0D | !A+ | Store to addr in A, inc A | A @ ! 1 A +!
> 0F | !A | Store to addr in A | A @ !
> 10 | com | Complement T | -1 XOR
> 11 | 2* | Left shift T, 0 to T0 | 2*
> 12 | 2/ | Right shift T, T20 to T19 | 2/
> 13 | +* | Add S to T if T0 is true | DUP 1 AND IF OVER + THEN
> 14 | -or | Exclusive-or S to T | XOR
> 15 | and | And S to T | AND
> 17 | + | Add S to T | +
> 18 | pop | Pop R, push to T | R>
> 19 | A | Push A to T | A @
> 1A | dup | Push T to T | DUP
> 1B | over | Push S to T | OVER
> 1C | push | Pop T, push to R | >R
> 1D | A! | Pop T to A | A !
> 1E | nop | Delay 2ns | NOP
> 1F | drop | Pop T | DROP

The 2ns delay ascribed to the nop instruction claims performance achieved using asynchronous logic design juju on specific silicon. Within the synchronous logic design discipline, expect nop to cost 1 clock cycle, rather than an absolute time.

Opcodes 04, 05, 07, 0E, and 16 are available for additions to the instruction set. The c18 B register is interesting.

- Myron

On Monday, April 18, 2022 at 10:00:16 PM UTC+10, Jan Coombs wrote:
> On Sun, 17 Apr 2022 22:07:00 -0700 (PDT)
> Wayne morellini <waynemo...@gmail.com> wrote:
>
> > Did Jeff ever reveal the instruction set for the F21? It would be good to examine his thinking on refining the instruction set, as he did a lot of simulation testing on it.
> F21 Microprocessor Preliminary specifications 9/98
>
> Code | Name | Description | Forth (with var named A)
> -----+------+------------------------------+--------------------------
> 00 | else | Unconditional jump | ELSE
> 01 | T0 | Jump is T0-19==0 | DUP IF
> 02 | call | Push PC+1 to R, jump | :
> 03 | C0 | Jump if T20 is false | CARRY? IF
> 06 | RET | Pop PC from R (return) | ;
> 08 | @R+ | Fetch from addr in R, inc R | R@ @ R> 1+ >R
> 09 | @A+ | Fetch from addr in A, inc A | A @ @ 1 A +!
> 0A | # | Fetch from PC+1, incr PC | LIT
> 0B | @A | Fetch from address in A | A @ @
> 0C | !R+ | Store to addr in R, inc R | R@ ! R> 1+ >R
> 0D | !A+ | Store to addr in A, inc A | A @ ! 1 A +!
> 0F | !A | Store to addr in A | A @ !
> 10 | com | Complement T | -1 XOR
> 11 | 2* | Left shift T, 0 to T0 | 2*
> 12 | 2/ | Right shift T, T20 to T19 | 2/
> 13 | +* | Add S to T if T0 is true | DUP 1 AND IF OVER + THEN
> 14 | -or | Exclusive-or S to T | XOR
> 15 | and | And S to T | AND
> 17 | + | Add S to T | +
> 18 | pop | Pop R, push to T | R>
> 19 | A | Push A to T | A @
> 1A | dup | Push T to T | DUP
> 1B | over | Push S to T | OVER
> 1C | push | Pop T, push to R | >R
> 1D | A! | Pop T to A | A !
> 1E | nop | Delay 2ns | NOP
> 1F | drop | Pop T | DROP
>
> abridged from:
> f21data__F21_PreliminarySpecifications_1998-09.pdf
>
>
> It would be good to be able to put this family of chips in date order,
> and note which ones made it to fully working silicon.
>
> The Mup21 was available in a PLCC44 package to buy in 1995. The video
> processor on the F21 is likely derived from the MuP21.
> > Actually, I have a though for an official open source Forth processor thread
> Good, even for commercial gain the open source route is probably most effective.
>
> Jan Coombs
> --
Jan. Thank you again for your useful contribution. I'm not with it today, but I can examine this against the mup21, and see what changes had been made, and if the instruction ordering has changed for some sort of circuitry design optimisation.

Jeffs chip is the last public large address space misc.

Is there a link to that document?

I finished a thread post before morning on the Forth Language Processor. But haven't reviewed it. I've had almost passing out like reaction to the vaccine from weeks ago today, so I think I'll skip reviewing it much and post it. Ting had already done p32 and p64 FPGA, so there are things out there people can look at for ideas.

Thanks.