I happened to have an email exchange with an APL friend of decades , Ray Cannon , which I copied to the SV-FIG mail group . It caused some stir illustrating the discontinuity in personal efficiency between GitHub , where tekys post code , and CoSy's use of its language on simple directories .
Sam Falvo , who gave an excellent SV-FIG talk on CoSy internals ( annotated ) , particularly the recursive reference counting memory management which evolves directly from Arthur Whitney , felt I was dissing tekys . Here are the emails , more or less in sequence .
On 15/03/2026 15:50, Bob
Armstrong wrote: | cc: SV-FIG
Ray ,Saw your comment on the Dyalog APL forum :
I am NOT a Github user and have no real wish to start using it. William Wang of Electronic Design some years ago suggested I put CoSy on GitHub . So I did .But it's ridiculously complex and limited compared to simply having a list language and an open directory .
A talk by John Peters at SV-FIG introduced me to cURL.se which has pretty universal Web IO so now I have some basic one-line
wwwGet and wwwPut verbs , and of course with CoSy vocabulary it's far more flexible than Git could hope to be . Git is for tekys to remain tekys in their own teky fiefdom , not for the humans simply wanting to get their own sh-- done and not forgotten who are my target . The main links I now post are :
Subject: Re: [svfig] Gagged on Git
Date: Mon, 16 Mar 2026 18:32:07 -0700
From: Samuel Falvo II <[email protected]>
To: Silicon Valley Forth Interest Group <[email protected]>
CC: Ray Cannon <[email protected]>, Bob Armstrong <[email protected]>
On Sun, Mar 15, 2026 at 8:50 AM Bob Armstrong via svfig <[email protected]> wrote:
William Wang of Electronic Design some years ago suggested I put CoSy on GitHub . So I did .
What was your purpose for doing so?
| 20260325.1622 | It really was for marketing ; that's where the teky market meets .
But it's ridiculously complex and limited compared to simply having a list language and an open directory .
Maybe. It depends on what you're trying to do. Git is a
*collaboration* tool. Its purpose is to allow many people,
especially if they don't directly know each other, to collaborate on a
single project. If you're not doing that, then yes, Git is
entirely the wrong tool for you.
A
talk by John Peters at SV-FIG introduced me to cURL.se which has pretty
universal Web IO so now I have some basic one-line wwwGet and
wwwPut verbs , and of course with CoSy vocabulary it's far more
flexible than Git could hope to be . Git is for tekys to remain
tekys in their own teky fiefdom , not for the humans simply
wanting to get
Umm, ..., no. You are free to hold your own opinions, but that stops when it becomes an open insult to others. The us-versus-them and No True Scotsman fallacy here is bald-faced and, frankly, offensive.
Either learn to use the tool properly, or move on. There is no
point in insulting an entire community for whom Git not only applies,
but without which you could not benefit from your own project.
Just how do you think cURL.se is built? (Answer: with Git.
See https://curl.se/dev/).
--
Samuel A. Falvo II
On 2026-03-16 04:34, Ray Cannon
wrote:
Bob,
Git(hub) is for sharing code to be modified and updated. I am writing my code for the fun of it, to prove to myself that I can. There is no benefit that can be gained by someone improving my code for WAV files, as one can simply "record" live music or use existing code such as MiDi files.
Useful code needs to be maintained, and that is best done by multiple people working together. Git is great for that, but far to complex for a single individual working alone.
Take Care
Ray
Subject: Re: [svfig] Gagged on Git
Date: Tue, 17 Mar 2026 00:26:24 -0700
From: Paul Hardy via svfig <[email protected]>
Reply-To: Silicon Valley Forth Interest Group <[email protected]>
To: Silicon Valley Forth Interest Group <[email protected]>
CC: Paul Hardy <[email protected]>, Ray Cannon <[email protected]>
On Mon, Mar 16, 2026 at 6:32 PM Samuel Falvo II via svfig wrote:
> Maybe. It depends on what you're trying to do. Git is
a *collaboration* tool. Its purpose is to allow many people,
especially if they don't directly know each other, to collaborate on a
single project. If you're not doing that, then yes, Git is
entirely the wrong tool for you.
>
I find Git's ability to handle branches of code useful even for a sole
developer (experimental branch that could break everything versus
stable branch, for example).
There is a free online introduction to Git at this Stanford link:
https://www-cs-students.stanford.edu/~blynn/gitmagic/
Chapter 2 introduces the basics.
Richard Silverman's _Git Pocket Guide_ from O'Reilly also provides a
concise introduction to Git, though probably dated (it's from 2013):
Tortoise Git is free GPL-licensed software that provides a GUI
interface for Windows, if that makes it easier to use and you have a
Windows system:
tortoisegit.org/
Have fun!
Paul Hardy
Subject: Re: [svfig] Gagged on Git
Date: Wed, 18 Mar 2026 09:31:53 -0600
From: Bob Armstrong <[email protected]>
To: Silicon Valley Forth Interest Group <[email protected]>
CC: John Peters <[email protected]>, Ray
Cannon <[email protected]>, Samuel Falvo II
<[email protected]>, Paul Hardy <[email protected]>
I have a lingering respiratory cold which cost me ~ 10 days . I want
to make a more thorough post in a monthly NL , along with comments
to Heartland WDC
Climate Conf I won't be able to get to following up on the
analysis of radiative equilibrium of a colored ball , like a planet
, discussed in NL202602 . The most
fundamental equations toward a planetary model are at CoSy.com/4thCoSy/Physics/general.f .
Note the Planck distribution is one line , but I'm hung up on CoSy's
lack of exponential output , which is a far more general annoyance .
I want to get across to the Climate crowd that ~278.7 has too
be understood as the most number from which any further computation
of Earth's temperature must be based .
| --
Anyway , getting back to Git .
I certainly had no thought of dissing Teky's with my comments on GitHub . I think , tho , as the examples below indicate , they do indicate the true paradigm break between having an open rich array language simply available at the tap of a key in your notes , versus a special purpose immutable menu hierarchy .
I'd like somebody to point out something GitHub does I can't do in a
line or 2 of CoSy .
For the nonce I'm just inserting my reply to Ray's reply .
Ray is now programming music in APL , and we go way back beyond some
adventures at the APL conference in Leningrad in 1992 .
Ray ,
I think Git is clumsy for anything . CoSy's note keeping environment , of course , evolved from APL's workspace , but with executions not messing with your notes . Any line is a potential command line to be edited as needed . And with lists-of-lists vocabulary to apply verbs to entire selections of files , menu structures are pointlessly prolix , idiosyncratic & time wasting .
It's trivial for someone to clone a copy of a directory for their own experimentation , but committing it to a centralized copy is a matter of mutual permissions , not interface .
Of course GitHub has built all sorts of add-on structures & services to the core .For example , why am I , or any normal person going to mess with the mass of GitHub when I can , for example , write
` https://cosy.com/4thCoSy/Code/CoSy/ wwwDir >t0> # |>| 55t0 5 _take fmttbl|
(
AcntFns.f | AltStackOps.f | BROWSER.txt | BasicHelpWords.html | CSauxstack.f
2024-05-08 | 2024-05-08 | 2024-05-08 | 2024-05-28 | 2024-05-08
15:41 | 15:41 | 15:42 | 11:22 | 15:42
)Or a line like
t0 ' dsc 'm .. ` .f conn at .. { t1 $ cL wwwGet # } 'm ,L flip fmttbl|
(
AcntFns.f | AltStackOps.f | CSauxstack.f
691 | 1565 | 2465
)| to select the [ FORTH ] scripts & list their sizes [ number of chars ]
Why , if you've got that power of array language at your fingertips are you going to cripple yourself with legacy tek , whether working as an individual , or having to submit requests to
` wwwPutsome files to a central committee for approval ?
And CoSy is open vocabulary all the way to the chip(s) .
| --
BTW , I may have mentioned her before , but are you familiar with the work of my childhood neighbor Laurie Spiegel ?
Bob A
On 2026-03-18 06:50, John Peters via
svfig wrote:
It does seem complicated to me a non-programmer, but it allows me
to work with two programmers, one in Japan and one here in San
Francisco to develop my Stock Watcher program.
Sent from my iPhone
On Mar 17, 2026, at 6:59 AM, Paul Hardy via svfig <[email protected]> wrote:On Tue, Mar 17, 2026 at 4:46 AM David Smith via svfig <[email protected]> wrote:
I support Bob's brief comments criticizing Git. I can imagine Git alternatives emerging from the Forth community. Do any exist already?
[ 20260325.1726 | Dave , open invitation to flesh it out in CoSy ]Several packages exist for Unix-like systems, including Cygwin for Windows. These come to mind:
* RCS, with its simple ci (check in) and co (check out) commands, from Unix's early days. It's about as simple as you can get. I used it in those early Unix days.
* CVS, which OpenBSD uses
* Mercurial: I experimented with this in the past because I was asked about using it on a project, but didn't stick with it. There are plugins for this in IDEs such as netbeans and Eclipse.
* Subversion: I used this before Git came along. Not anywhere near as powerful as Git and simpler as a consequence.
There's also System V Unix SCCS, but that has fallen by the wayside.Paul Hardy
`( Fact SunSpecs )` Dv@
>T0> fmttbl|
r |
696000000.0000
Dist | 149597900000.0000
temp |
5778.0000
| /\ | exponential formated output somehow has never yet gotten
implemented in Reva FORTH , so I apologize for sucky
formatting . The same is true for ` genericising math verbs . CoSy
executes directly in x86 FORTH so lotta ` fs needed just to label
` Float operations . See
.
| \/ | Solid Angle , steradians , from
radius and distance .
T0 `( r Dist )` '
v@ 'R ^2f ' %f ./ Pi *f ,/ >t0> fmtF18 ,/
|>| 0.000068001322294938 Converting steradians to portion of entire celestial
sphere :
t0
Pi `f 4 *f %f .. f1. $ -f %f fmtF18 ,/
|>| 0.000005411402573131
That 5.4 millionths of the
celestial sphere around us at about 6000K provides all the energy to
our planet .
| making those calcs into verbs
| Solid Angle from radius and distance :
: R_D>strad ^2f ' %f
./ Pi *f ;
| steradians to portion of sphere
: sdrad>PofSphere Pi 4.
_f *f %f .. f1. $ -f %f ;
So now the computation can be written :
T0 `( r Dist )` ' v@
'R R_D>strad sdrad>PofSphere >t1> `f 1.e6 *f ,/
|>| 5.4114
and the portion covered by the Sun and its complement are
(
0.000005411402573131
0.999994588597426868
)
So now the gray body temperature in our orbit , can be calculated ,
using a value of 3K for dark space , by raising the temps to the 4th
power , adding them , and taking 4th root . Note the Stefan Boltzmann
Constant , : sb 5.6704e-8 _f ;
, falls out of the round trip from
temp to temp , giving :
f( 5778.0 3.0
)f `f 4 ^f t2 *f +/ `f 0.25 ^f |>| 278.68
Matching Grok's calculation .
Now I will go thru the computation for the temperature of a radiantly
heated uniformly colored ball
, as in the image below .
As I learned as a grade school nerd during the Ike admin , and Grok
explains , it is NOT explained
simply by the overall reflectivity , Bond
albedo , of the ball .
0 Comments
Log in to join the conversation.No comments yet. Be the first to share your thoughts.