In my *Server* table I have (most definitions removed)

(opened) {nnfolder:archive}
...
(opened) {nnimap:pi2-a.local}
(opened) {nnimap:pi2-b.local}
...
(opened) {nndraft:}
(opened) {nnml:}

and wanted to add two entries for

{nntp:pi2-a.local}
{nntp:pi2-b.local}

and then strange things happened. The table seemed to behave strangely
when trying to move/reorder them and [e]diting those new nntp: ones
interfered with the nnimap: definitions with the same key(?).
Additionally even the motion in the table seemed to jump strangely
around when manipulating the added entries at the bottom.

(nnimap "pi2-a.local"
;; ^
;; Key, tag, nickname, ... does that have an official name?
...)

I reverted all changes and thought about using the list of secondary
methods in `.gnus` instead[0], but I may have been already to tired to
do that correctly and/or this runs into the same problem when keys
collide.

Looking into `.newsrc.eld`, I see

(setq gnus-server-alist
'(
...
("pi2-a.local" nnimap "pi2-a.local" ...)
("pi2-b.local" nnimap "pi2-b.local" ...)
...
)
)

and get the idea that adding

("pi2-a.local" nntp "pi2-a.local" ...)
("pi2-b.local" nntp "pi2-b.local" ...)

may stumble over using the same key in the first component. So far this
only is a shot into the dark.

I did not have that problem when I wired my old two NNTP server guinea
pigs into GNUS, which were named completely different (key and
hostnames) because I used my naming scheme for (Tor) hidden services.

After a longer break, I'll retry this with including the protocol in the
keys, which may be bearable because it just would look like

imap.hostname(.domain)
nntp.hostname(.domain)
...

unless a better idea ferments over the span of that break.

____________

[0]: Nevertheless I'm thinking I should switch to defining my secondary
methods in `.gnus` because the editable snippets in the *Server*
table do not allow comments and I do miss being able to comment
those entries really a lot.
--
ERROR: SIGNATURE NOT FOUND.
I'll DJT you! I'll DJT you! I'll DJT you! I'll DJT you!
I'll DJT you! I'll DJT you! I'll DJT you! I'll DJT you!
PLEASE RESTART TOURRETTE-LLM ASAP.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

In my *Server* table I have (most definitions removed)

(opened) {nnfolder:archive}
...
(opened) {nnimap:pi2-a.local}
(opened) {nnimap:pi2-b.local}
...
(opened) {nndraft:}
(opened) {nnml:}

and wanted to add two entries for

{nntp:pi2-a.local}
{nntp:pi2-b.local}

and then strange things happened. [...]

That's the main flaw of Gnus. It's not predictable. Like sed. It's
not intuitive. Gnus at least is only one, while sed too many.
Nevertheless, Gnus is the most pleasant to use, so we insist.

I think it's important to ask why it's the most pleasant. I believe
it's because it lives in the GNU EMACS. We could replace Gnus
completely if our operating system were more like a Lisp Machine from
the 80s. Does this make sense?
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

It's not predictable. Like sed. It's not intuitive.

Both

Stand-up Maths
How on Earth does ^.?$|^(..+?)\1+$ produce primes?
<https://www.youtube.com/watch?v=5vbk0TwkokM>

and

laserbat
A sed script that draws mandelbrot sed
<https://github.com/laserbat/mandelbrot.sed>

have some smile over `sed` potential, but IMO the second one still has a
glitch in the calculation or uses a nonstandard way to "colour" the
result. I'm not really fluent with `sed` details, so I decided not to
dig deeper.
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

That's the main flaw of Gnus. It's not predictable. Like sed. It's
not intuitive. Gnus at least is only one, while sed too many.
Nevertheless, Gnus is the most pleasant to use, so we insist.

I think it's important to ask why it's the most pleasant. I believe
it's because it lives in the GNU EMACS.

We can do much more in EmacsOS than we do, so while we are only seeing
the nose of the swimming Ursus Maritimus (to avoid the iceberg theme),
we that way sure only see a tiny fraction of the complexity too.
Despite massively deviating from the UI of other EmacsOS apps, GNUS
still interacts well with its neighbours. Probably that is what keeps
me using GNUS.

We could replace Gnus completely if our operating system were more
like a Lisp Machine from the 80s. Does this make sense?

I heavily doubt that Lisp is the right tool for everything.

Read the above as: I have no idea and I am too gnarly to switch away
from Unix.
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

It's not predictable. Like sed. It's not intuitive.

Both

Stand-up Maths
How on Earth does ^.?$|^(..+?)\1+$ produce primes?
<https://www.youtube.com/watch?v=5vbk0TwkokM>

and

laserbat
A sed script that draws mandelbrot sed
<https://github.com/laserbat/mandelbrot.sed>

have some smile over `sed` potential, but IMO the second one still has a glitch in the calculation or uses a nonstandard way to "colour" the
result. I'm not really fluent with `sed` details, so I decided not to
dig deeper.

These are puzzles. I usually don't spend any time on these things.
It's kinda like security vulnerabilities. You should find them on the
software you're working on, but it senseless to try to exploit them
because it's a lot of work that has no future.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

[...]

We could replace Gnus completely if our operating system were more
like a Lisp Machine from the 80s. Does this make sense?

I heavily doubt that Lisp is the right tool for everything.

Read the above as: I have no idea and I am too gnarly to switch away
from Unix.

I didn't quite mean that Lisp would be the language, but that type of
features there were on those machines. Self-documented, debugged while
it was running, read the source as you use the software and so on.

But I like to see Lisp in everything. It's a marvelous language and I
think the s-exp is a big reason. So easy to manipulated in chunks---paredit-mode I mean.
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

These are puzzles.

s/puzzles/coding or fun/

I really can like that!
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

These are puzzles.

s/puzzles/coding or fun/

I really can like that!

Okay, here's one for you. Using POSIX sh and POSIX echo alone, make
echo print -n. Like this:

$ cmdline...
-n
$

I'm sure you know how to use echo.
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

Okay, here's one for you. Using POSIX sh and POSIX echo alone, make
echo print -n. Like this:

$ cmdline...
-n
$

I'm sure you know how to use echo.

`echo` can be very diverse on the Nixens and one better takes nothing
for granted. I've no idea how correct `yash` sticks to Posix, but it
was worth a try:

$ yash -o posixlycorrect
$ echo -n
-n
$ ^D

I'm not a fan of *SH any more. Today I hate the risks of "..${STUFF}.."
in strings more than I did some decades ago. For me SH has degraded to
just another language for fun and puzzles.

Just some weeks ago I finally tried to get `${!name}` (see BASH) and
arrays in plain SH using `eval`. Sure it can be done, that probably is
one of the reasons `eval` was included in SH, but it definitely does not
get a readability award. Coding in such a painful language[0] should go extinct in our 13th millennium[1].

_____________________

[0]: I include every classic shell and their over-featured siblings
there! There is no need to have the command line frontend to the
system and the scripting language(s) use the same syntax.

[1]: Holocene calendar
<https://en.wikipedia.org/wiki/Holocene_calendar>
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

I use GNUS, and I am interest on how you made your footnotes start on 0.

Please share how :-) .
--- Synchronet 3.20a-Linux NewsLink 1.2

keyboardan <keyboardan@tilde.club> wrote:

I use GNUS, and I am interest on how you made your footnotes start on 0.

Please share how :-) .

GNUS can do footnotes on its own?

I just write them manually.

The only time I see "footnote" mentioned in GNUS' single HTML page
manual is as the heading of the footnotes section at its end.

Maybe some other Emacs functionality[1] would allow footnotes in messages
or texts in general, but that's just not what I use.

You can dig for "style" and "numeric" in

emacs/lisp/mail/footnote.el (emacs' git repo)

and if your attention span is bigger than mine, might find an easy patch
for it.


Footnotes:
[1] Aaaaaah: footnote-mode and `M-x footnote-add-footnote`. So far I
only had tried Orgmode's footnotes once.
--
/"\ This virus has not been found by antivirus software. /"\
\!/ To replicate it needs your help: \!/
_|_ Please copy it to your signature to ensure its survival. _|_
/ V \ _____ \\o o// \o/ _____ THANKS!! _____ \o/ \\o o// _____ / V \
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

keyboardan <keyboardan@tilde.club> wrote:

I use GNUS, and I am interest on how you made your footnotes start on 0.

Please share how :-) .

GNUS can do footnotes on its own?

Yes, indeed.

I just write them manually.

The only time I see "footnote" mentioned in GNUS' single HTML page
manual is as the heading of the footnotes section at its end.

Maybe some other Emacs functionality[1] would allow footnotes in messages
or texts in general, but that's just not what I use.

Footnotes:
[1] Aaaaaah: footnote-mode and `M-x footnote-add-footnote`. So far I
only had tried Orgmode's footnotes once.

That is the it!

You can dig for "style" and "numeric" in

emacs/lisp/mail/footnote.el (emacs' git repo)

and if your attention span is bigger than mine, might find an easy patch
for it.

I have patched it, but I don't know if I took the easy path. Although it
works now ;-) With the following:

```
(with-eval-after-load "footnote"
(defun footnote--insert-footnote (arg)
"Insert a footnote numbered ARG, at (point)."
(push-mark)
(let ((ptr (footnote--insert-numbered-footnote arg t)))
(footnote--goto-char-point-max)
(if (footnote--goto-first)
(save-restriction
(when footnote-narrow-to-footnotes-when-editing
(footnote--narrow-to-footnotes))
(footnote-goto-footnote (1- arg)) ; evil, FIXME (less evil now)
(message "Inserting footnote %d" arg)
(or (eq arg 0)
(when (re-search-forward
(if footnote-spaced-footnotes
"\n\n"
(concat "\n" (footnote--current-regexp)))
nil t)
(beginning-of-line)
t)
(footnote--goto-char-point-max)
(footnote--goto-first)))
(unless (looking-at "^$")
(insert "\n"))
(when (eobp)
(insert "\n"))
(unless (string-equal footnote-section-tag "")
(insert footnote-section-tag "\n")))
(let ((text (footnote--insert-numbered-footnote arg nil)))
(footnote--insert-markers arg text ptr))))

(defun footnote--make-hole ()
"Make room in the alist for a new footnote at point.
Return the footnote number to use."
(save-excursion
(let (rc)
(dolist (alist-elem footnote--markers-alist)
(when (<= (point) (caddr alist-elem))
(unless rc
(setq rc (car alist-elem)))
(save-excursion
(message "Renumbering from %s to %s"
(footnote--index-to-string (car alist-elem))
(footnote--index-to-string
(1+ (car alist-elem))))
(footnote--renumber (1+ (car alist-elem))
alist-elem))))
(or rc
(1+ (or (caar (last footnote--markers-alist)) (1- 0)))))))

(defun footnote-renumber-footnotes ()
"Renumber footnotes, starting from 0."
(interactive "*")
(save-excursion
(let ((i 0))
(dolist (alist-elem footnote--markers-alist)
(footnote--renumber i alist-elem)
(setq i (1+ i)))))))
```

Happy holidays!
--- Synchronet 3.20a-Linux NewsLink 1.2

keyboardan <keyboardan@tilde.club> writes:

I use GNUS, and I am interest on how you made your footnotes start on 0.

Why would you want it to start at 0? ``First footnote'' matches with
the number 1, not zero. :)
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

But I like to see Lisp in everything. It's a marvelous language and I
think the s-exp is a big reason. So easy to manipulated in chunks---paredit-mode I mean.

I still have some problems with deeply nested notations, not only in the flavours of Lisp.

E.g.

f(g(h(x))) # BASIC, C, Pascal, ...

(f (g (h x))) # Lisp family

forces me to read down to h or even x to grok what's going on, while
reading

x | h | g | f

is more step by step and at least easier for my brain.

Sometimes even the "write-only languages" DC and Forth look clearer:

x h g f # Forth

x lhx lgx lfx # DC

Maybe someday I find something in Scheme that fits this control flow
better or grok how to create such via hygienic macros. My to do list is
far too long and maybe I'll never get there.


Have you seen PureData?

My dream would be a front-end like PD to "draw" the program and then to
export that to the desired target language.

Thinking about code should not be tied to specific languages until one
starts to implement it for a target system.

Maybe even the view as graph or puzzle pieces should not stay the only
way. Converting an algorithm into different representations depending on
which bug one is hunting may help a lot.
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

But I like to see Lisp in everything. It's a marvelous language and I
think the s-exp is a big reason. So easy to manipulated in
chunks---paredit-mode I mean.

I still have some problems with deeply nested notations, not only in the flavours of Lisp.

E.g.

f(g(h(x))) # BASIC, C, Pascal, ...

(f (g (h x))) # Lisp family

forces me to read down to h or even x to grok what's going on, while
reading

x | h | g | f

is more step by step and at least easier for my brain.

Sometimes even the "write-only languages" DC and Forth look clearer:

x h g f # Forth

x lhx lgx lfx # DC

Maybe someday I find something in Scheme that fits this control flow
better or grok how to create such via hygienic macros. My to do list is
far too long and maybe I'll never get there.

Have you seen PureData?

I have never seen PureData (or PD). Although it seems that PD can also
be a programming language.

I wonder if it is possible to program macros, and to do
meta-programming, in PD. If not, I support LISP everywhere.

I already support LISP everywhere, and I am open to other ideas.

Maybe McCarthy was trying to solve this reading problem of the LISP
code, with M-expressions. But most people decided that would be okay to
write and read S-expressions.

My dream would be a front-end like PD to "draw" the program and then to export that to the desired target language.

Thinking about code should not be tied to specific languages until one
starts to implement it for a target system.

Should it be tied to the PD programming language?

Maybe even the view as graph or puzzle pieces should not stay the only
way. Converting an algorithm into different representations depending on which bug one is hunting may help a lot.

It seems you will like low-code software development. One low-code software/plataform/language is OutSystems[0]. OutSystems is not Free
Software.


[0] https://www.outsystems.com/
--- Synchronet 3.20a-Linux NewsLink 1.2

keyboardan <keyboardan@tilde.club> wrote:

yeti <yeti@tilde.institute> writes:

Thinking about code should not be tied to specific languages until one
starts to implement it for a target system.

Should it be tied to the PD programming language?

Using PD is like using a CAD program.
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

keyboardan <keyboardan@tilde.club> wrote:

yeti <yeti@tilde.institute> writes:

Thinking about code should not be tied to specific languages until one
starts to implement it for a target system.

Should it be tied to the PD programming language?

Using PD is like using a CAD program.

Okay, didn't know that.

For me is more productive to type than to use the mouse/touchpad for
most stuff, including for programming software.
--- Synchronet 3.20a-Linux NewsLink 1.2

keyboardan <keyboardan@tilde.club> wrote:

yeti <yeti@tilde.institute> writes:

keyboardan <keyboardan@tilde.club> wrote:

yeti <yeti@tilde.institute> writes:

Thinking about code should not be tied to specific languages until one >>>> starts to implement it for a target system.

Should it be tied to the PD programming language?

Using PD is like using a CAD program.

Okay, didn't know that.

For me is more productive to type than to use the mouse/touchpad for
most stuff, including for programming software.

I prefer the keyboard too, but something PD like even could be usable on
a text screen via arrow keys. That's a secondary problem to solve.
Even in ASCII diagrams such a representation may be of use and I'm not
thinking of only wanting that one, as import and export to other
notations should be available too.

But this will stay SciFi.

*ping!*

NEXT TOPIC PLEASE!
--
Tweets are Xcretions now.
--- Synchronet 3.20a-Linux NewsLink 1.2

Patricia Ferreira <pferreira@example.com> wrote:

Why would you want it to start at 0? ``First footnote'' matches with
the number 1, not zero. :)

Why?

Uints start at zero. Why waste the 0 when just needing individual tags?
After all there are different ways to link footnotes. You probably
remember the * or the dagger, so even easily[ॐ] distinguishable[💣] emojis[👀] would do that job.

Look into footnote.el for some alternatives.

I just had some binge physicsing with a video series with the 1st
episode numbered as "#000". Why not?

IIRC I saw footnotes with short multi-letter tags somewhere. In times
of klick[RET] to jump, the only reason against that might be that the
link to the footnote should™ not occupy too much space.

____________

[ॐ]: Ok, that may be a bit too ॐmmmptimistic!

[👀]: I've some doubts, but today's kids seem to love carving new
hieroglyphs into bitstone. Have you seen emojicode[🤯]?

[🤯]: Emojicode is an open-source, full-blown programming language
consisting of emojis.
<https://www.emojicode.org/>

[💣]: Ok, I'll stop it here. I'm not a kid. For me all these new
hieroglyphs look like near to indistinguishable fly shit on my
screen. But if jumping between the footnote anchor and its text
just would be like C-c C-c, only Emacs would need to be able to
find the right target. We probably all are not reading our news
after printing it, so even that could work.

[RET]: Or was it ^N?

™: Nah! This was not meant to be a footnote, but let's add one as
precaution.
--
1. Hitchhiker 17: (25) "Shhh!" said Zaphod. "There's absolutely nothing
to be worried about."
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

Patricia Ferreira <pferreira@example.com> wrote:

Why would you want it to start at 0? ``First footnote'' matches with
the number 1, not zero. :)

Why?

Uints start at zero. Why waste the 0 when just needing individual tags? After all there are different ways to link footnotes. You probably
remember the * or the dagger, so even easily[ॐ] distinguishable[💣] emojis[👀] would do that job.

Look into footnote.el for some alternatives.

I just had some binge physicsing with a video series with the 1st
episode numbered as "#000". Why not?

IIRC I saw footnotes with short multi-letter tags somewhere. In times
of klick[RET] to jump, the only reason against that might be that the
link to the footnote should™ not occupy too much space.

____________

[ॐ]: Ok, that may be a bit too ॐmmmptimistic!

[👀]: I've some doubts, but today's kids seem to love carving new
hieroglyphs into bitstone. Have you seen emojicode[🤯]?

[🤯]: Emojicode is an open-source, full-blown programming language
consisting of emojis.
<https://www.emojicode.org/>

[💣]: Ok, I'll stop it here. I'm not a kid. For me all these new
hieroglyphs look like near to indistinguishable fly shit on my
screen. But if jumping between the footnote anchor and its text
just would be like C-c C-c, only Emacs would need to be able to
find the right target. We probably all are not reading our news
after printing it, so even that could work.

[RET]: Or was it ^N?

™: Nah! This was not meant to be a footnote, but let's add one as
precaution.

"First" is an ordinal number and ordinal numbers follow the ancient
Roman method of counting that starts on "1", because there was no "0" in ancient Roman numbers.

But there came Fibonacci and introduced the "0" before the "1". And,
nowadays, we use the Fibonacci numbers. So, today, the ordinal numbers
should start on "0", but people don't do that. "first" should be "0th". "second" be "1th", "third" "2th", etc..

Also, today we should start counting on "0", since "zero" is today's
first number. "one" is the second number. Only intelligent people do
start counting on 0.

Yeah, the Math you learned in school is illogical, even badly insane.

(Some of the few places, in today's Math, where "0" is honoured, is in Euclidean Space (geometrical cardinal points). Acceleration also has
"0", but that is called Physics, and today, Physics has been separated
from Math)
--- Synchronet 3.20a-Linux NewsLink 1.2

keyboardan <keyboardan@tilde.club> wrote:

Yeah, the Math you learned in school is illogical, even badly insane.

Even the Peano axioms seem to have been updated somewhen.

<https://en.wikipedia.org/wiki/Peano_axioms>

I learned them in a variant starting at 1 and stepping by one and always thought that were strange.

At the university I had professors sticking to 1 as start of N, others preferred 0. Insanely both groups used the same symbol N and to add the tching-bang to that insanity used N with index 0 to mean the other interpretation, so one index 0 meant add the 0, the other one meant
remove the 0.

And I some days even had lessons with members of both interpretations
and I still think that was not helpful.

POWER TO THE ZERO!
--
I do not bite, I just want to play.
--- Synchronet 3.20a-Linux NewsLink 1.2

yeti <yeti@tilde.institute> writes:

keyboardan <keyboardan@tilde.club> wrote:

Yeah, the Math you learned in school is illogical, even badly insane.

Even the Peano axioms seem to have been updated somewhen.

<https://en.wikipedia.org/wiki/Peano_axioms>

I learned them in a variant starting at 1 and stepping by one and always thought that were strange.

At the university I had professors sticking to 1 as start of N, others preferred 0. Insanely both groups used the same symbol N and to add the tching-bang to that insanity used N with index 0 to mean the other interpretation, so one index 0 meant add the 0, the other one meant
remove the 0.

And I some days even had lessons with members of both interpretations
and I still think that was not helpful.

Good to know that I am not alone in thinking that the Math taught to us
at school is illogical.

POWER TO THE ZERO!

:-)
--- Synchronet 3.20a-Linux NewsLink 1.2