2 Apr
2025
2 Apr
'25
3:05 a.m.
So, I am working on another fun project on my trusty old TRS-80's.
I need the ability to do multi-line functions which TRS-80 BASIC
does not support. But in hopes of finding something where someone
added that capability in the distant past (there were lots of features
added by various people over time) I decided to just try a Google
Search. As we all know Google now begins results with an AI generated
answer to your question. This is what I got:
--------------------------------------------------------------------------
In TRS-80 BASIC, you can create multi-line functions (subroutines)
using the DEF FN statement, allowing you to define a function that
can perform a specific task and be called from different parts of
your program.
Here's how it works:
DEF FN Statement:
You use the DEF FN statement to define a function, followed by the
function name (e.g., FNMyFunction), a set of parentheses for any
arguments, and then the code that the function will execute.
Example: 10 DEF FNC(X,Y) = X*X + Y*Y defines a function named FNC
that takes two arguments (X and Y) and returns the sum of their squares.
Function Call:
To use the function, you simply call it by its name, passing in the
necessary arguments.
Example: 20 PRINT FNC(3,4) would print the value returned by the FNC
function when called with arguments 3 and 4, which is 25.
Multi-line Functions:
You can extend the DEF FN statement to span multiple lines, allowing
for more complex logic within the function.
Example:
Code
10 DEF FNC(X,Y)
20 LET TEMP = X * X
30 LET TEMP = TEMP + Y * Y
40 DEF FN END
50 PRINT FNC(3,4)
This example defines a function that calculates the sum of squares
of two numbers, but the calculation is broken down into multiple
lines for readability.
Key Points:
Line Numbers:
TRS-80 BASIC relies on line numbers, so make sure to use them
appropriately when defining multi-line functions.
Function Name:
Function names must start with FN followed by a valid identifier
(e.g., FNC, FNMyFunction).
DEF FN END:
You need to use DEF FN END to mark the end of the function
definition, especially when using multi-line functions.
Scope:
Functions are defined within the scope of the program and are
accessible from anywhere in the program.
-----------------------------------------------------------------------
Which is both wrong and totally absurd.
And I see things all the time about letting AI write your code for you.
Does anyone actually trust AI to write code?
bill
2 Apr
2 Apr
3:17 a.m.
And now I have to clean the coffee out of my keyboard....
I asked the question with slightly different wording
and got a totally different explanation with even
more absurd examples. I think I'll ask it to write
me a PL/M Compiler in TRS-80 BASIC. :-)
bill
3:32 a.m.
One shortcoming, that has come up before, is that it is taught to NOT say,
"I don't know"
Somebody asked ChatGPT for the last eight digits of PI, and it gave a
sequence of 8 digits. But, somebody else asked, "Is there a last digit of
PI?", and it gave a reasonably correct answer.
Perhaps, ask it to do it in Level I BASIC, and fit in 4K, . . .
--
Grumpy Ol' Fred cisin(a)xenosoft.com
On Tue, 1 Apr 2025, Bill Gunshannon via cctalk wrote:
And now I have to clean the coffee out of my keyboard....
I asked the question with slightly different wording
and got a totally different explanation with even
more absurd examples. I think I'll ask it to write
me a PL/M Compiler in TRS-80 BASIC. :-)
bill
5:22 a.m.
On 4/1/25 18:32, Fred Cisin via cctalk wrote:
One shortcoming, that has come up before, is that it
is taught to NOT
say, "I don't know"
Apparently, some think that within 10 years, AI will replace doctors and
teachers:
https://people.com/bill-gates-ai-will-replace-doctors-teachers-in-next-10-y…
Undoubtedly, it will all be powered by nuclear fusion.
I dunno, I'd be a bit leery if my medical treatment would be dictated by
an AI, particularly by one that didn't ever say "I don't know..."
And back in the dark ages, I studied MEDLARS when it ran on 7090s and
tape drives.
--Chuck
3 Apr
3 Apr
11:21 p.m.
New subject: Why I am not worried about AI taking over the world!
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk wrote:
Apparently, some think that within 10 years, AI will
replace doctors and
teachers:
It's one thing to say that AI methods have been shown to give better
diagnoses in some situations than the humans, it's quite another to
suggest we won't need doctors. In my neck of the woods, even doctors
have been greatly supplanted by physician assistants.
I'm as skeptical as they come, and I have plenty of criticisms of AI
as implemented today, but I've also been confounded to discover how
useful it can be when programming. It's often better than googling and
copy-and-pasting from Stackoverflow.
I've generally used Copilot as it comes with M365. In a recent Perl tool
that needed a few dozen regular expressions, it was very handy, providing
not only correct regexs but also snippets of the surrounding code.
In another task where I wanted to quickly generate a text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
I've heard similar puzzlement from a very experienced programmer
friend who told me about the way his IDE would now suggest code
to him while he was typing, and he thought the suggested code was
good and very similar if not identical to what he was about to write.
- John
11:37 p.m.
On 4/3/25 14:21, John Foust via cctalk wrote:
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk wrote:
In another task where I wanted to quickly generate a
text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
Interesting. I got no help at all from ChatGPT when working through
some IBM 7090 code.
--Chuck
11:50 p.m.
On 04/03/2025 5:37 PM EDT Chuck Guzis via cctalk
<cctalk(a)classiccmp.org> wrote:
On 4/3/25 14:21, John Foust via cctalk wrote:
My experience has been much like John's. In my company of 10,000 ish employees, the
entire company is "strongly encouraged" to use various "AI" tools made
available. The encouragement comes directly from the CEO (and all his minions echo it)
during his monthly video addresses.
So, all of us who write code have copilot licenses and they actually track how much we use
it (can you say layoff? I knew you could.) What I have found is that the short
suggestions of a half dozen lines or more are often pretty good. It's the bad ones I
have to fight the system to NOT take that eat up the time. I usually could have just
written those half dozen lines faster. But sometimes it helps a lot. Apparently it
tracks names and kind find the correct variable or constant faster than I could. A couple
of coworkers have given a good description: good for boilerplate code.
As for chatgpt not doing well, I suspect that is largely due to it being a general purpose
"question answerer" rather than a tool specifically for programming. I know of
at least one person who did, however, get hired based mostly on code written by chatgpt.
But he didn't last long.
Will
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk
wrote:
In another task where I wanted to quickly
generate a text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
Interesting. I got no help at all from ChatGPT when working through
some IBM 7090 code.
--Chuck 
4 Apr
4 Apr
12:52 a.m.
On Thu, Apr 3, 2025, 3:51 PM Will Cooke via cctalk <cctalk(a)classiccmp.org>
wrote:
That's what I've found too. Generate the boilerplate for me: not very
creative but usuall 90+% right for code I'll be rototilling anyway.
Small code: usually ok. Generate about 10-50 lines it's ok. But much more
than that and the errors are too high to be worth it. Debugging time
becomes much higher than the just write it from scratch time. Small
snippets are also easier to catch errors on too.
It's not terrible. On par with autocomplete emacs macros I've sometimes
written but with less effort.
But it's not great either, at least for the code I write.
Warner
As for chatgpt not doing well, I suspect that is largely due to it being a
On 04/03/2025 5:37 PM EDT Chuck Guzis via cctalk
<cctalk(a)classiccmp.org>
wrote:
On 4/3/25 14:21, John Foust via cctalk wrote:
My experience has been much like John's. In my company of 10,000 ish
employees, the entire company is "strongly encouraged" to use various
"AI"
tools made available. The encouragement comes directly from the CEO (and
all his minions echo it) during his monthly video addresses.
So, all of us who write code have copilot licenses and they actually track
how much we use it (can you say layoff? I knew you could.) What I have
found is that the short suggestions of a half dozen lines or more are often
pretty good. It's the bad ones I have to fight the system to NOT take that
eat up the time. I usually could have just written those half dozen lines
faster. But sometimes it helps a lot. Apparently it tracks names and kind
find the correct variable or constant faster than I could. A couple of
coworkers have given a good description: good for boilerplate code.
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk
wrote:
In another task where I wanted to quickly
generate a text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
Interesting. I got no help at all from ChatGPT when working through
some IBM 7090 code.
--Chuck
general purpose "question answerer" rather
than a tool specifically for
programming. I know of at least one person who did, however, get hired
based mostly on code written by chatgpt. But he didn't last long.
Will
1:14 a.m.
On 4/3/25 16:37, Chuck Guzis via cctalk wrote:
On 4/3/25 14:21, John Foust via cctalk wrote:
Well, that likely dates WAY before the large language model
has seen.
Jon
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk
wrote:
In another task where I wanted to quickly generate a text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
Interesting. I got no help at all from ChatGPT when working through
some IBM 7090 code. 
4:29 a.m.
On 4/3/25 16:37, Chuck Guzis via cctalk wrote:
On 4/3/25 14:21, John Foust via cctalk wrote:
I have a friend who's an EE and taking advanced math and statistics
courses at his university. He's middle aged, a brilliant engineer, and
blessed with at least some common sense.
He uses ChatGPT and a couple of other free AI services to write a lot of
code for his classwork (with his instructors' blessing), and we've
talked a lot about using that effectively.
The short version is that ChatGPT works well for writing functions and
snippets - small, mostly linear pieces - but not complex processes. And
more importantly, you have to be able to write what you're asking for,
AND VALIDATE the output.
Finally, he says it's a lot like a web search. The query is key to
getting correct results. The ability to diagnose incorrect results and
modify the query is a determining factor.
Doc
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk
wrote:
In another task where I wanted to quickly
generate a text-based
profile of a computer's specifications and components, it was very
helpful in writing a bunch of PowerShell. Did it hallucinate?
Perhaps once out of several dozen things I asked it, and in that
case, it was easy to see that the two clauses of the if-then were
the same.
Interesting. I got no help at all from ChatGPT when working through
some IBM 7090 code.
3:19 p.m.
New subject: Why I am not worried about AI taking over the world!
At 09:29 PM 4/3/2025, Doc Shipley via cctalk wrote:
The short version is that ChatGPT works well for
writing functions and snippets - small, mostly linear pieces - but not complex processes.
And more importantly, you have to be able to write what you're asking for, AND
VALIDATE the output.
Sounds like he needs to write longer, more complex queries. It's perfectly
possible to hand it nine pages of description of what you want.
To keep this slightly more on-topic, if there is demand for AI to generate
or explain code of old dialects of BASIC, who out there is training
a model on old BASIC source code?
- John
5 Apr
5 Apr
4:37 a.m.
I've played with using ChatGPT to write code for older things (quickbasic,
1990's C++, stuff like that). In my experience it gets confused and gives
you snippets that has features from later, modern languages. I'm far from
an expert but it feels to me like it gets confused during processing about
what to use as a source when giving it's answer.
So based on that, I'd say that the training wouldn't help a whole lot.
On Fri, Apr 4, 2025 at 10:06 AM John Foust via cctalk <cctalk(a)classiccmp.org>
wrote:
At 09:29 PM 4/3/2025, Doc Shipley via cctalk wrote:
The short version is that ChatGPT works well for
writing functions and
snippets - small, mostly linear pieces - but not complex processes. And
more importantly, you have to be able to write what you're asking for, AND
VALIDATE the output.
Sounds like he needs to write longer, more complex queries. It's perfectly
possible to hand it nine pages of description of what you want.
To keep this slightly more on-topic, if there is demand for AI to generate
or explain code of old dialects of BASIC, who out there is training
a model on old BASIC source code?
- John
5:15 a.m.
On 2025-04-04 8:37 p.m., Michael Huff via cctalk wrote:
I've played with using ChatGPT to write code for
older things (quickbasic,
1990's C++, stuff like that). In my experience it gets confused and gives
you snippets that has features from later, modern languages. I'm far from
an expert but it feels to me like it gets confused during processing about
what to use as a source when giving it's answer.
So based on that, I'd say that the training wouldn't help a whole lot.
I thought FORTRAN IV was the portable programing language.
They have talked about having smarter high level programing languages
for years. Has that gotten anywhere?
5:44 a.m.
On 4/4/25 20:15, ben via cctalk wrote:
I thought FORTRAN IV was the portable programing
language.
They have talked about having smarter high level programing languages
for years. Has that gotten anywhere?
Back in the heyday of FORTRAN, the universe of architectures was a lot
more diverse and foreign to languages such as C. Non-binary (decimal)
numeric representations, Binary ones' complement math, vendor-unique
character sets (it's why FORTRAN (and COBOL) is written using a very
small common character set)--and lack of character-handling operations,
as well as Boolean functions.
I think that the original PALASM was written in "portable" FORTRAN. One
of the more common ways to start a "portable" program was to READ a card
with the character set punched onto it in A1 format. You manipulated
characters by referring to their positions in the alpha A1 array.
--Chuck
6 Apr
6 Apr
1:41 a.m.
In those days, FORTRAN IV was handy as a "portable" language just
because it was the one language (other than COBOL) available
*everywhere*. It could be used as a sort of "high level assembler" too.
My first job out of college was working on a large database system. It
was written using a pre-processor (think of RATFOR) and a bit of
per-platform assembler for things that couldn't be done in standard
FORTRAN (e.g. random access files, OS calls). It was a real, high
performance database system that ran on CDC, UNIVAC, IBM 370, Xerox
SIGMA, PDP-10/20, and VAX. Each new platform usually turned up some
minor quirks. We always said there were no portable programs, only
programs that had been ported.
Gary
On 4/4/25 23:44, Chuck Guzis via cctalk wrote:
On 4/4/25 20:15, ben via cctalk wrote:
I thought FORTRAN IV was the portable programing
language.
They have talked about having smarter high level programing languages
for years. Has that gotten anywhere?
Back in the heyday of FORTRAN, the universe of architectures was a lot
more diverse and foreign to languages such as C. Non-binary (decimal)
numeric representations, Binary ones' complement math, vendor-unique
character sets (it's why FORTRAN (and COBOL) is written using a very
small common character set)--and lack of character-handling operations,
as well as Boolean functions.
I think that the original PALASM was written in "portable" FORTRAN. One
of the more common ways to start a "portable" program was to READ a card
with the character set punched onto it in A1 format. You manipulated
characters by referring to their positions in the alpha A1 array.
--Chuck
2:23 a.m.
In those days, FORTRAN IV was handy as a
"portable" language just because it
was the one language (other than COBOL) available *everywhere*. It could be
used as a sort of "high level assembler" too.
"FORTRAN --"the infantile disorder"--, by now nearly 20 years old, is
hopelessly inadequate for whatever computer application you have in mind
today: it is now too clumsy, too risky, and too expensive to use."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
"The use of COBOL cripples the mind; its teaching should, therefore, be
regarded as a criminal offence."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
Dijkstra had things to say about many languages.
In 1975, . . .
Fortran77 or Fortran90 did fix a few things, but not necessarily the right
ones.
Who was who said that FORTRAN was "more portable than syphilis"?
7 Apr
7 Apr
5 p.m.
a friend asked me whether ChatGPT designed and planned
DOGE ?
The DOGE Cryptocurrency predates the LLM by quite a long time.
The LLM stuff is wild, if you haven't tried the AI stuff it's a game
changer for sure. I can see it leading to job losses. It know it already
has.
It's helped me develop software for a board that allows me to read data
out of a running Williams Pinball Controller (Pinball machine CPU board
from the 90s.) I'm able to SSH in to the Pi side and read any memory
addresses while the pinball machine is running.
Early on I was trying to get it to write me a converter to go from MIDI
file to the Atari Music Composer cartdrige format. While it didn't get it
right, the fact that it kind of got close was wild. Could of worked with
it some more and it probably would have gotten it right.
I watched a friend rapidly make a script to parse a really gnarly web
page. I was complaining that I was having difficulty narrowing down the
data from all the div tags and he delivers a working (and tiny) script in
3 minutes.
Don't underestimate this stuff.
It could and show upend our medical system. Odds are with a good data set
behind it -- it would beat our doctors and could allow everyone to receive
better/more medical analysis. But the AMA probably won't allow it -- to
protect their profits.
It's easy to feel you are un-replacable but honestly most people are. I
remember a pizza chain in Virginia Beach. IT contractor had set up a
system using the BBS FrontDoor software that nightly would take the data
from the point of sale systems at 6 or 7 stores and send it to central
location. He had compiled all the batch files into .COM or .EXE files so
they couldn't be modified, and had set passwords and had some kind of back
door. Had himself locked in for good gig with this odd setup. Low and
behold a friend of mine who had experience with the BBS software, slid in
and replaced it all in a flash one night. Could see the other guy
connecting trying to use his back door but nada. From that point I always
thought "Never get too cocky."
--
: Ethan O'Toole
5:23 p.m.
New subject: Why I am not worried about AI taking over the world!
At 10:00 AM 4/7/2025, Ethan O'Toole via cctalk wrote:
Don't underestimate this stuff.
Thank you for the confirmation.
Nay-saying is always easy. I'd rather hear from smart people who
can tell me whether or how a tool is useful now, as well as how it
might enhance projects in the future.
Why couldn't a present or future AI help you translate from one BASIC
dialect to another, if that's what you want to do? If you want a tool
to do that, how could you help make it possible? How can AI tools
help with disassembly? How might they help you search and understand
and get answers from all the files in your digital packratting?
For the record, my first paying programming job in '81 was translating
between dialects of BASIC.
- John
9 Apr
9 Apr
1:19 a.m.
Nay-saying is always easy. I'd rather hear from
smart people who
can tell me whether or how a tool is useful now, as well as how it
might enhance projects in the future.
Why couldn't a present or future AI help you translate from one BASIC
dialect to another, if that's what you want to do? If you want a tool
to do that, how could you help make it possible? How can AI tools
help with disassembly? How might they help you search and understand
and get answers from all the files in your digital packratting?
For the record, my first paying programming job in '81 was translating
between dialects of BASIC.
- John
It possibly could! Try it with ChatGPT and see if it works.
For an example of how bad ass the AI stuff is.... A friend and I are
supposed to supply a bunch of restored pinball and arcades to a bar
restaurant arcade type thing. Vendors are asking around $10,000 for a
change machine that can take credit card payment and dispense tokens.
The people putting the place together asked if machines can be had used. I
tell them, for dollar bills easy but not the modern credit card ones. Can
I make one, they ask. I start digging around looking for info on those
terminals and how transaction setups work for them to judge if it would be
easy to add a subsystem to an older change machine to accept card payments
and dispense tokens. Developer only portals that require signups, fake
sites trying to harvest data and sell you their services, sigh. Google
used to be better.
Hit up ChatGPT. A few specifically asked questions and it craps out
example code for telling the exact terminal I asked about the amount you
wish to charge and picking up status. Very cool, and more than I was
looking for. Confirms that the card info staying on the PCI compliant CC
terminal and never crossing to my system should meet PCI compliance. Asked
it more questions about payment gateways, processors, and is it possible
to share those accounts between popular touchscreen point of sale
terminals. The thing answered a lot and gave a lot of specific examples of
various brands of equipment and if you can share the accounts. All in like
5 minutes. I could get it to crank out the code for the end user UI as
well as code for arduino to manage communications to the 90s era change
machine easily. It just reduced a bunch of web searches to 5 minutes with
very specific examples. Some risk it could be wrong, but nothing major.
It's wild, give it a try. ChatGPT is free. Throw it some odd stuff, and
give it quite a bit of detail in your question. I just asked it to write
me code for IMSAI 8080 to bounce LED back and forth and it is cranking it
out with comments in the code. It tells me about the front panel switches,
but if I have a monitor rom it's easier. Asked me if I want instructions
on toggling it in or data for a rom monitor even.
At VCF East someone griped that these GPT things are providing data that
is stolen from other people. Be it the artwork or the code or whatever. I
kind of fail to see the difference in training it on a large data set
versus some kid learning to play piano by playing the Thexder theme or
Arbands books for brass or whatever. We all build on what have learned
from other people's examples.
It's neat.
--
: Ethan O'Toole
7:39 p.m.
On 4/8/2025 7:19 PM, Ethan O'Toole via cctalk wrote:
All this talk of ChatGPT caught my interest. Es[ecially the
part about it writing code to actually fulfill a task. So, I asked it
to do some coding. It told me what needed to be done to complete
the task. (I already knew all that!) And spit out a few generic
blocks of code. Not really useful. And in some cases, just like
with google, wrong like suggesting the use of CHARACTER variables in
FORTRAN IV.
I'll continue to play with it but my initial reactions still stand.
No sign of intelligence and not ready for prime-time.
bill
Nay-saying is
always easy. I'd rather hear from smart people who
can tell me whether or how a tool is useful now, as well as how it
might enhance projects in the future.
Why couldn't a present or future AI help you translate from one BASIC
dialect to another, if that's what you want to do? If you want a tool
to do that, how could you help make it possible? How can AI tools
help with disassembly? How might they help you search and understand
and get answers from all the files in your digital packratting?
For the record, my first paying programming job in '81 was translating
between dialects of BASIC.
- John
It possibly could! Try it with ChatGPT and see if it works.
For an example of how bad ass the AI stuff is.... A friend and I are
supposed to supply a bunch of restored pinball and arcades to a bar
restaurant arcade type thing. Vendors are asking around $10,000 for a
change machine that can take credit card payment and dispense tokens.
The people putting the place together asked if machines can be had used.
I tell them, for dollar bills easy but not the modern credit card ones.
Can I make one, they ask. I start digging around looking for info on
those terminals and how transaction setups work for them to judge if it
would be easy to add a subsystem to an older change machine to accept
card payments and dispense tokens. Developer only portals that require
signups, fake sites trying to harvest data and sell you their services,
sigh. Google used to be better.
Hit up ChatGPT. A few specifically asked questions and it craps out
example code for telling the exact terminal I asked about the amount you
wish to charge and picking up status. Very cool, and more than I was
looking for. Confirms that the card info staying on the PCI compliant CC
terminal and never crossing to my system should meet PCI compliance.
Asked it more questions about payment gateways, processors, and is it
possible to share those accounts between popular touchscreen point of
sale terminals. The thing answered a lot and gave a lot of specific
examples of various brands of equipment and if you can share the
accounts. All in like 5 minutes. I could get it to crank out the code
for the end user UI as well as code for arduino to manage communications
to the 90s era change machine easily. It just reduced a bunch of web
searches to 5 minutes with very specific examples. Some risk it could be
wrong, but nothing major.
It's wild, give it a try. ChatGPT is free. Throw it some odd stuff, and
give it quite a bit of detail in your question. I just asked it to write
me code for IMSAI 8080 to bounce LED back and forth and it is cranking
it out with comments in the code. It tells me about the front panel
switches, but if I have a monitor rom it's easier. Asked me if I want
instructions on toggling it in or data for a rom monitor even.
At VCF East someone griped that these GPT things are providing data that
is stolen from other people. Be it the artwork or the code or whatever.
I kind of fail to see the difference in training it on a large data set
versus some kid learning to play piano by playing the Thexder theme or
Arbands books for brass or whatever. We all build on what have learned
from other people's examples.
It's neat.
6 Apr
6 Apr
4:31 a.m.
On 4/5/2025 8:23 PM, Fred Cisin via cctalk wrote:
And it was comments like those that convinced me Dyjkstra was an idiot
and I have never placed any value on anything he ever said.
bill
In those days,
FORTRAN IV was handy as a "portable" language just
because it was the one language (other than COBOL) available
*everywhere*. It could be used as a sort of "high level assembler" too.
"FORTRAN --"the infantile disorder"--, by now nearly 20 years old, is
hopelessly inadequate for whatever computer application you have in mind
today: it is now too clumsy, too risky, and too expensive to use."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
"The use of COBOL cripples the mind; its teaching should, therefore, be
regarded as a criminal offence."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
Dijkstra had things to say about many languages.
In 1975, . . .
Fortran77 or Fortran90 did fix a few things, but not necessarily the
right ones.
1:50 p.m.
Dijkstra was my GOTO man for anything to do with computing! :-)
Nigel
On 2025-04-05 22:31, Bill Gunshannon via cctalk wrote:
On 4/5/2025 8:23 PM, Fred Cisin via cctalk wrote:
And it was comments like those that convinced me Dyjkstra was an idiot
and I have never placed any value on anything he ever said.
bill
--
Nigel Johnson, MSc., MIEEE, MCSE VE3ID/G4AJQ/VA3MCU
Amateur Radio, the origin of the open-source concept!
Skype: TILBURY2591
In those
days, FORTRAN IV was handy as a "portable" language just
because it was the one language (other than COBOL) available
*everywhere*. It could be used as a sort of "high level assembler"
too.
"FORTRAN --"the infantile disorder"--, by now nearly 20 years old, is
hopelessly inadequate for whatever computer application you have in
mind today: it is now too clumsy, too risky, and too expensive to use."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
"The use of COBOL cripples the mind; its teaching should, therefore,
be regarded as a criminal offence."
- "How do we tell truths that might hurt?"
- Edsger W.Dijkstra, 18 June 1975
Dijkstra had things to say about many languages.
In 1975, . . .
Fortran77 or Fortran90 did fix a few things, but not necessarily the
right ones.
5 Apr
5 Apr
4:03 p.m.
On 4/4/2025 11:15 PM, ben via cctalk wrote:
On 2025-04-04 8:37 p.m., Michael Huff via cctalk
wrote:
I never thought of Fortran of being portable by design. It may
be fairly portable just because of the simplicity of the language.
I'll let you know, I am about to try moving a program that ran on
TOPS-10 and IBM MTS to Microsoft Fortran-80. :-)
The truly portable language was COBOL. :-)
bill
I've played with using ChatGPT to write code
for older things
(quickbasic,
1990's C++, stuff like that). In my experience it gets confused and gives
you snippets that has features from later, modern languages. I'm far from
an expert but it feels to me like it gets confused during processing
about
what to use as a source when giving it's answer.
So based on that, I'd say that the training wouldn't help a whole lot.
I thought FORTRAN IV was the portable programing language.
They have talked about having smarter high level programing languages
for years. Has that gotten anywhere?
5:08 p.m.
On 4/5/25 07:03, Bill Gunshannon via cctalk wrote:
I never thought of Fortran of being portable by
design. It may
be fairly portable just because of the simplicity of the language.
I'll let you know, I am about to try moving a program that ran on
TOPS-10 and IBM MTS to Microsoft Fortran-80. :-)
I did something similar some 50 years ago. Took the PDP-10 "Colossal
Cave/Adventure" source to CDC 6000 FORTRAN running under SCOPE.
The part that involved the most labor was conversion of the source code
tape that used the TOPS-10 convention of 5 7-bit ASCII in a 36 bit word
to CDC 6 bit Display code. Conversion of the FORTRAN dialect was fairly
trivial; the most effort involved re-implementing the SAVE command (CDC
SCOPE did not have a simple method for saving core images and restarting
them). The result ran almost from the first try.
The truly portable language was COBOL. :-)
50 years ago, COBOL (and FORTRAN) standardization was a loose affair.
Many smaller systems simply lacked COBOL (was there a 1401 COBOL?).
Language features were removed as the machine architecture dictated.
For example, contrast 1960s IBM S/360 USA "BASIC" FORTRAN with, say
UNIVAC 1108 FORTRAN-V. Going from the small (no logical IF, etc.) IBM
dialect to the UNIVAC one would have been pretty simple; the other
direction, not so much.
I recall that DOS/360 COBOL as released lacked ISAM--the statements
could be compiled, but no code was generated. Instead, IBM provided a
bunch of library routines that were invoked with the "ENTER LINKAGE"
statement.
--Chuck
--Chuck
7:53 p.m.
New subject: AI taking over the world!
Perhaps it is time to look at AI, 50 years ago.
Byte vol 3 jan - brains of men and machines.
Have we gotten farther in better computer tech?
Ben.
GO ANALOG - GET YOUR BRAIN IN A BOTTLE TODAY 99 cent special,
SEE 'EVIL MINIONS R US'.
4 Apr
4 Apr
1:12 a.m.
On 4/3/25 16:21, John Foust via cctalk wrote:
At 10:22 PM 4/1/2025, Chuck Guzis via cctalk wrote:
I am also skeptical. I am a minor developer on the LinuxCNC
project and somebody suggested getting chatGPT to come up
with a script for setting the I?O side of that up. Since
this is a VERY niche project, I didn't expect it to do
much. I was amazed that it created a script that a cursory
scan indicated it might actually work!
Our IT guy at work swears by chatGPT. He was trying to
debug some issues with VirtualBox setup after a kernel
upgrade on my Linux desktop. (My fix is to disable ALL
updates). Anyway, he was going crazy, and wasted about TWO
DAYS trying to fix it! I went straight to the Oracle web
site and downloaded the update file, and then typed the
commands exactly as shown on their update page, and it
worked perfectly! I am guessing that chatGPT was making a
response based on previous versions of the software, and
those were just wrong for the latest version.
So, I've seen both.
As for PAs, my regular doctor was out of town, I had a rash
on my side and went to an urgent care place, the PA there
totally misdiagnosed it, and it got worse. Then, I went to
a dermatologist, he took a sample of scrapings off the rash,
diagnosed it correctly, and the prescription solved it. Not
sure my regular doctor would have been able to diagnose it
either.
Jon
Apparently, some think that within 10 years, AI
will replace doctors and
teachers:
It's one thing to say that AI methods have been shown to give
better
diagnoses in some situations than the humans, it's quite another to
suggest we won't need doctors. In my neck of the woods, even doctors
have been greatly supplanted by physician assistants.
I'm as skeptical as they come, and I have plenty of criticisms of AI
as implemented today, but I've also been confounded to discover how
useful it can be when programming. It's often better than googling and
copy-and-pasting from Stackoverflow.
1:23 a.m.
On 4/3/25 16:12, Jon Elson via cctalk wrote:
As for PAs, my regular doctor was out of town, I had a
rash on my side
and went to an urgent care place, the PA there totally misdiagnosed it,
and it got worse. Then, I went to a dermatologist, he took a sample of
scrapings off the rash, diagnosed it correctly, and the prescription
solved it. Not sure my regular doctor would have been able to diagnose
it either.
Like every other human endeavor, there are good and not-so-good
professionals. I've gotten somewhat better attention from a DO than an
MD, but that could be the luck of the draw.
Being a cancer patient with a rare sarcoma, I've seen a *lot* of doctors
lately--with more to come.
--Chuck
1:36 a.m.
I’ve not had good luck with any PA’s so I won’t see them.
Their diagnosis is supposed to be reviewed/approved by a real Doctor, but I don’t think
that happens like it’s supposed to.
Sent from my iPhone
On Apr 3, 2025, at 16:23, Chuck Guzis via cctalk
<cctalk(a)classiccmp.org> wrote:
On 4/3/25 16:12, Jon Elson via cctalk wrote:
As for PAs, my regular doctor was out of town, I
had a rash on my side
and went to an urgent care place, the PA there totally misdiagnosed it,
and it got worse. Then, I went to a dermatologist, he took a sample of
scrapings off the rash, diagnosed it correctly, and the prescription
solved it. Not sure my regular doctor would have been able to diagnose
it either.
Like every other human endeavor, there are good and not-so-good
professionals. I've gotten somewhat better attention from a DO than an
MD, but that could be the luck of the draw.
Being a cancer patient with a rare sarcoma, I've seen a *lot* of doctors
lately--with more to come.
--Chuck
7:55 p.m.
As for PAs, my regular doctor was out of town, I had a
rash on my side and went
to an urgent care place, the PA there totally misdiagnosed it, and it got
worse. Then, I went to a dermatologist, he took a sample of scrapings off the
rash, diagnosed it correctly, and the prescription solved it. Not sure my
regular doctor would have been able to diagnose it either.
<physician>
Admittedly, the fact he scraped it and got a sample is likely *why* he got it
right. When in doubt about a rash, get a chunk of it. I haven't been in general
practice in over a decade but I never referred until I had a pathology result.
On the other hand, general medicine and family medicine practices have so
little time for visits now. I don't miss that grind.
I don't think AI will do a good job of taking biopsies ...
</physician>
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser(a)floodgap.com
-- Premature optimization is the root of all evil. -- Donald Knuth ------------
5 Apr
5 Apr
3:24 a.m.
On 4/4/25 12:55, Cameron Kaiser via cctalk wrote:
Right, I don't think
my internist would have had the
reagents or whatever on hand to identify the cause of the
rash, but I think he would have recognized it as an
infection. The PA gave me some steroid cream which was
totally the wrong thing to do.
As for PAs, my
regular doctor was out of town, I had a rash on my side and went
to an urgent care place, the PA there totally misdiagnosed it, and it got
worse. Then, I went to a dermatologist, he took a sample of scrapings off the
rash, diagnosed it correctly, and the prescription solved it. Not sure my
regular doctor would have been able to diagnose it either.
<physician>
Admittedly, the fact he scraped it and got a sample is likely *why* he got it
right. When in doubt about a rash, get a chunk of it. I haven't been in general
practice in over a decade but I never referred until I had a pathology result.
On the other hand, general medicine and family medicine practices have so
little time for visits now. I don't miss that grind. I don't think AI will do a good job of taking
biopsies ...
</physician>
Right, some thing need hands or eyes right on the scene.
Jon
2 Apr
2 Apr
4:06 p.m.
On Tuesday, April 1st, 2025 at 21:32, Fred Cisin via cctalk <cctalk(a)classiccmp.org>
wrote:
One shortcoming, that has come up before, is that it
is taught to NOT say,
"I don't know"
Heh, I worked with a guy like that once. Unfortunately, he was in sales, so field
engineering then got to explain that what he'd sold them was actually not feasible.
Thanks,
Jonathan
6:35 p.m.
On 4/2/2025 10:06 AM, Jonathan Chapman via cctalk wrote:
On Tuesday, April 1st, 2025 at 21:32, Fred Cisin via
cctalk <cctalk(a)classiccmp.org> wrote:
In my days as a beltway bandit I was very popular as a pre-sales system
engineer. At least, with the customers. I never let a salesman try to
sell something I knew engineering could never deliver.
bill
One shortcoming, that has come up before, is that
it is taught to NOT say,
"I don't know"
Heh, I worked with a guy like that once. Unfortunately, he was in sales, so field
engineering then got to explain that what he'd sold them was actually not feasible.
4:26 a.m.
I have a book about writing AI code on a Tandy computer from 1981
On Tue, Apr 1, 2025 at 10:01 PM Bill Gunshannon via cctalk <
cctalk(a)classiccmp.org> wrote:
So, I am working on another fun project on my trusty old TRS-80's.
I need the ability to do multi-line functions which TRS-80 BASIC
does not support. But in hopes of finding something where someone
added that capability in the distant past (there were lots of features
added by various people over time) I decided to just try a Google
Search. As we all know Google now begins results with an AI generated
answer to your question. This is what I got:
--------------------------------------------------------------------------
In TRS-80 BASIC, you can create multi-line functions (subroutines)
using the DEF FN statement, allowing you to define a function that
can perform a specific task and be called from different parts of
your program.
Here's how it works:
DEF FN Statement:
You use the DEF FN statement to define a function, followed by the
function name (e.g., FNMyFunction), a set of parentheses for any
arguments, and then the code that the function will execute.
Example: 10 DEF FNC(X,Y) = X*X + Y*Y defines a function named FNC
that takes two arguments (X and Y) and returns the sum of their squares.
Function Call:
To use the function, you simply call it by its name, passing in the
necessary arguments.
Example: 20 PRINT FNC(3,4) would print the value returned by the FNC
function when called with arguments 3 and 4, which is 25.
Multi-line Functions:
You can extend the DEF FN statement to span multiple lines, allowing
for more complex logic within the function.
Example:
Code
10 DEF FNC(X,Y)
20 LET TEMP = X * X
30 LET TEMP = TEMP + Y * Y
40 DEF FN END
50 PRINT FNC(3,4)
This example defines a function that calculates the sum of squares
of two numbers, but the calculation is broken down into multiple
lines for readability.
Key Points:
Line Numbers:
TRS-80 BASIC relies on line numbers, so make sure to use them
appropriately when defining multi-line functions.
Function Name:
Function names must start with FN followed by a valid identifier
(e.g., FNC, FNMyFunction).
DEF FN END:
You need to use DEF FN END to mark the end of the function
definition, especially when using multi-line functions.
Scope:
Functions are defined within the scope of the program and are
accessible from anywhere in the program.
-----------------------------------------------------------------------
Which is both wrong and totally absurd.
And I see things all the time about letting AI write your code for you.
Does anyone actually trust AI to write code?
bill
5:13 a.m.
I have a book about writing AI code on a Tandy
computer from 1981
Commodore did one of these too. Very primitive, but I liked it at the time.
--
------------------------------------ personal: http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com * ckaiser(a)floodgap.com
-- Complexity kills. -- Ray Ozzie ---------------------------------------------
1:19 p.m.
I'm not worried about AI taking over the world.
What worries me is politicians and commercial management believing the hype
and turning the world over to AI before it works properly.
On Wed, Apr 2, 2025 at 8:36 AM Cameron Kaiser via cctalk <
cctalk(a)classiccmp.org> wrote:
I have a book about writing AI code on a Tandy
computer from 1981
Commodore did one of these too. Very primitive, but I liked it at the time.
--
------------------------------------ personal:
http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com *
ckaiser(a)floodgap.com
-- Complexity kills. -- Ray Ozzie
---------------------------------------------
8 Apr
8 Apr
2:10 p.m.
On Wed, 2 Apr 2025 at 12:36, Adrian Godwin via cctalk
<cctalk(a)classiccmp.org> wrote:
I'm not worried about AI taking over the world.
Good.
What worries me is politicians and commercial
management believing the hype
Yes, that is a real problem.
and turning the world over to AI before it works
properly.
It's not _going_ to work properly. There is no such thing as "AI" yet
and it is an unsolved question if there ever will be.
What is 100% clear and apparent _if_ one is willing to take a proper
deep look is that LLM bots are not in any way intelligent _and never
will be_.
They really are nothing more than very very clever autocomplete
engines and _can't_ be anything more ever.
It is not actually all that hard to understand how they work and what
they can't do.
This is my favourite short easy explainer, from of all places the
_Financial Times_ --
https://ig.ft.com/generative-ai/
This is a longer and much more in-depth one from the programmer of
_Mathematica_ --
https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-d…
Saying this baldly violates the new religion of the Silicon Valley
believers, but LLM bots at not AI and they never will be. They fail
both the "A" and the "I".
--
Liam Proven ~ Profile: https://about.me/liamproven
Email: lproven(a)cix.co.uk ~ gMail/gTalk/FB: lproven(a)gmail.com
Twitter/LinkedIn: lproven ~ Skype: liamproven
IoM: (+44) 7624 227612: UK: (+44) 7939-087884
Czech [+ WhatsApp/Telegram/Signal]: (+420) 702-829-053
9:46 p.m.
On Tue, Apr 8, 2025 at 1:10 PM Liam Proven via cctalk <cctalk(a)classiccmp.org>
wrote:
I don't think that the lack of a usable candidate is proof that there
never will be. I do agree that it won't be LLM.
and turning
the world over to AI before it works properly.
It's not _going_ to work properly. There is no such thing as "AI" yet
and it is an unsolved question if there ever will be.
Saying this baldly violates the new religion of the Silicon Valley
believers, but LLM bots at not AI and they never will be. They fail
both the "A" and the "I".
Imho they're artificial in the way artificial turf is. There are a limited
number of applications in which it can stand in for actual turf. In that
sense I don't think the term 'AI', which changes it's meaning with each
generation is terribly useful other than to marketing people. Bill Hicks
had some thoughts about them. I prefer the term 'machine intelligence' to
describe something that can compete with human minds. It's further away
than I can imagine but maybe not inconceivable.
2 Apr
2 Apr
4:37 p.m.
I did a talk at this year's Trenton Computer Festival on the history of AI.
On Wed, Apr 2, 2025 at 3:36 AM Cameron Kaiser via cctalk <
cctalk(a)classiccmp.org> wrote:
I have a book about writing AI code on a Tandy
computer from 1981
Commodore did one of these too. Very primitive, but I liked it at the time.
--
------------------------------------ personal:
http://www.cameronkaiser.com/ --
Cameron Kaiser * Floodgap Systems * www.floodgap.com *
ckaiser(a)floodgap.com
-- Complexity kills. -- Ray Ozzie
---------------------------------------------
5:59 a.m.
I did some programming on the TRS-80 Model 4 back in the 80's so it has been a long
time.
I did add some functionality using machine code. I had a screen that displayed options to
the user but took a long time to print it. So the first time it printed it dump the screen
to memory. The every other time it was needed it move the info from the data to the
screen.
For a function, I used a subroutine. For example:
100 X=20 'argument of function
110 GOSUB 1000 'calculate y=myfunction(x)
120 R=A+B*Y
130 continue program
1000 REM this is the myfunction, input x, returns y
1010 function statements
1020
1030
1040
1050
1080 Y= .... REM Result of function
1090 RETURN
Don't know if that if what you were looking for. Obviously not as clean as
R=A*MYFUNCTION(X)
----- Original Message -----
From: "Bill Gunshannon via cctalk" <cctalk(a)classiccmp.org>
To: cctalk(a)classiccmp.org
Cc: "Bill Gunshannon" <bill.gunshannon(a)hotmail.com>
Sent: Tuesday, April 1, 2025 9:05:56 PM
Subject: [cctalk] Why I am not worried about AI taking over the world!
So, I am working on another fun project on my trusty old TRS-80's.
I need the ability to do multi-line functions which TRS-80 BASIC
does not support. But in hopes of finding something where someone
added that capability in the distant past (there were lots of features
added by various people over time) I decided to just try a Google
Search. As we all know Google now begins results with an AI generated
answer to your question. This is what I got:
--------------------------------------------------------------------------
In TRS-80 BASIC, you can create multi-line functions (subroutines)
using the DEF FN statement, allowing you to define a function that
can perform a specific task and be called from different parts of
your program.
Here's how it works:
DEF FN Statement:
You use the DEF FN statement to define a function, followed by the
function name (e.g., FNMyFunction), a set of parentheses for any
arguments, and then the code that the function will execute.
Example: 10 DEF FNC(X,Y) = X*X + Y*Y defines a function named FNC
that takes two arguments (X and Y) and returns the sum of their squares.
Function Call:
To use the function, you simply call it by its name, passing in the
necessary arguments.
Example: 20 PRINT FNC(3,4) would print the value returned by the FNC
function when called with arguments 3 and 4, which is 25.
Multi-line Functions:
You can extend the DEF FN statement to span multiple lines, allowing
for more complex logic within the function.
Example:
Code
10 DEF FNC(X,Y)
20 LET TEMP = X * X
30 LET TEMP = TEMP + Y * Y
40 DEF FN END
50 PRINT FNC(3,4)
This example defines a function that calculates the sum of squares
of two numbers, but the calculation is broken down into multiple
lines for readability.
Key Points:
Line Numbers:
TRS-80 BASIC relies on line numbers, so make sure to use them
appropriately when defining multi-line functions.
Function Name:
Function names must start with FN followed by a valid identifier
(e.g., FNC, FNMyFunction).
DEF FN END:
You need to use DEF FN END to mark the end of the function
definition, especially when using multi-line functions.
Scope:
Functions are defined within the scope of the program and are
accessible from anywhere in the program.
-----------------------------------------------------------------------
Which is both wrong and totally absurd.
And I see things all the time about letting AI write your code for you.
Does anyone actually trust AI to write code?
bill
6:46 p.m.
On 4/1/2025 11:59 PM, bobh(a)tds.net wrote:
I did some programming on the TRS-80 Model 4 back in
the 80's so it has been a long time.
I did add some functionality using machine code. I had a screen that displayed options to
the user but took a long time to print it. So the first time it printed it dump the screen
to memory. The every other time it was needed it move the info from the data to the
screen.
For a function, I used a subroutine. For example:
100 X=20 'argument of function
110 GOSUB 1000 'calculate y=myfunction(x)
120 R=A+B*Y
130 continue program
1000 REM this is the myfunction, input x, returns y
1010 function statements
1020
1030
1040
1050
1080 Y= .... REM Result of function
1090 RETURN
Don't know if that if what you were looking for. Obviously not as clean as
R=A*MYFUNCTION(X)
I knew (and tried a little bit to see how hard it would be) that method.
But the catch is at this stage I am trying to convert a program from
NorthStar BASIC to TRS-80 BASIC. And all the multi-line functions are
already there but, obviously, need to be replaced. Along with a lot
of other conversions. Like the fact that there really are no strings
in NorthStar BASIC. They are actually character arrays and bring with
that some quirks that TRS-80 BASIC doesn't have.
Because this is just a starting point for my long term project (really
long term - I started this back when the TRS-80 was still mainline :-)
I am thinking I will give up efficiency and just put the function code
inline everywhere there is a call to that function.
bill
6:46 p.m.
On Wed, Apr 2, 2025 at 3:01 AM Bill Gunshannon via cctalk
<cctalk(a)classiccmp.org> wrote:
So, I am working on another fun project on my trusty old TRS-80's.
I need the ability to do multi-line functions which TRS-80 BASIC
does not support.
Actually there was an offical BASIC (meaning sold by Tandy/Radio
Shack) for one family of TRS-80 machines that did support multi-line
(named) procedures.
Not strictly functons as there was no return value but as you can pass
parameters by reference you can return values that way.
And line numbers were optional.
It's Microware's BASIC-09 under OS-9 on the CoCo of course. Quite the
finest 8-bit BASIC that I have ever used.
-tony
6:54 p.m.
On 4/2/2025 12:46 PM, Tony Duell wrote:
On Wed, Apr 2, 2025 at 3:01 AM Bill Gunshannon via
cctalk
<cctalk(a)classiccmp.org> wrote:
Not meaning to be pedantic, but, OS-9 ran on the COCO which
was never called a TRS-80 that I know of. The COCO had the
usual version of Microsoft BASIC found on the TRS-80's and
while Radio Shack sold Basic09 it was not their product, it
was from Microware just like OS-9. I have used it and this
project may eventually drift in that direction after the TRS-80
part is done, as portability is one of my goals. I have at
least one and maybe more, I forget, Basic09 programs on
Rosetta Code.
bill
So, I am working on another fun project on my trusty old TRS-80's.
I need the ability to do multi-line functions which TRS-80 BASIC
does not support.
Actually there was an offical BASIC (meaning sold by Tandy/Radio
Shack) for one family of TRS-80 machines that did support multi-line
(named) procedures.
Not strictly functons as there was no return value but as you can pass
parameters by reference you can return values that way.
And line numbers were optional.
It's Microware's BASIC-09 under OS-9 on the CoCo of course. Quite the
finest 8-bit BASIC that I have ever used.
7:07 p.m.
On Wed, Apr 2, 2025 at 5:54 PM Bill Gunshannon via cctalk
<cctalk(a)classiccmp.org> wrote:
Not meaning to be pedantic, but, OS-9 ran on the COCO
which
was never called a TRS-80 that I know of. The COCO had the
The service manual for the CoCo 2 is titled :
'Radio Shack Service Manual
TRS-80 (R)
Color Computer 2'
The CoCo 3 manual does not say TRS-80 on it through
-tony
3 Apr
3 Apr
8:43 p.m.
Tony Duell says:
The CoCo 3 manual does not say TRS-80 on it through
By then the computers decided that "TRS-80"/"Radio Shack" was a
liability, so called new computers "Tandy". Thus Tandy 1000, Tandy 6000
(succeeding the TRS-80 Model 16B and Model 12), and Tandy Color Computer 3.
--
geo:37.783333,-122.416667
8 Apr
8 Apr
2:11 p.m.
On Wed, 2 Apr 2025 at 17:54, Bill Gunshannon via cctalk
<cctalk(a)classiccmp.org> wrote:
the COCO which
was never called a TRS-80 that I know of.
Yes it was.
Note the title of this page:
https://en.wikipedia.org/wiki/TRS-80_Color_Computer
Here's a pic:
https://en.wikipedia.org/wiki/TRS-80_Color_Computer#/media/File:TRS-80_Colo…
--
Liam Proven ~ Profile: https://about.me/liamproven
Email: lproven(a)cix.co.uk ~ gMail/gTalk/FB: lproven(a)gmail.com
Twitter/LinkedIn: lproven ~ Skype: liamproven
IoM: (+44) 7624 227612: UK: (+44) 7939-087884
Czech [+ WhatsApp/Telegram/Signal]: (+420) 702-829-053
16
days inactive
23
days old
45 comments
22 participants
participants (22)
-
Adrian Godwin -
ben -
Bill Degnan -
Bill Gunshannon -
bobh＠tds.net -
Cameron Kaiser -
Chuck Guzis -
Doc Shipley -
Ethan O'Toole -
Fred Cisin -
Gary Grebus -
John Foust -
Jon Elson -
Jonathan Chapman -
Liam Proven -
Michael Huff -
Nigel Johnson Ham -
Tony Duell -
Warner Losh -
Wayne S -
wrcooke＠wrcooke.net -
Yeechang Lee