Thinks I hate from #GNU and a bit from #Unix/#X11:

- #Emacs Calc (core package) depending on nonfree (not
GPL compatible) Gnuplot for plots. They should rewrite
the output for GNU #Plotutils and enhance the second
to have 3D plot support.

https://spdx.org/licenses/gnuplot.html

Distributing Gnuplot forks as a patch
against the main one -mandatory- makes Gnuplot's
code unshareable with other projects, period.

- #Texinfo (official GNU documentation format)
depending on #Texlive where tons of stuff in the official
release it's nonfree. Again, that should be a core feature
for #texinfo, and not depend on #tex at all.

https://wiki.parabola.nu/TeXLive_freedom_verification

- #Xedit might look as a crude X11 text editor but it
supports vi command AND it has an embedded Lisp
interpreter (compatible with a bit of CL/core Elisp
code, enough to do Math). It's 2026 and still doesn't
support neither UTF8 nor Unicode fonts. Trying to
enforce them under Xaw3D (LD_PRELOAD) will crash XEdit.
That editor would have lots of support if it supported them.
It's far lighter than Emacs and with the bundled Lisp
you might do wonders.

Back to Groff/Troff.
At least under #Hyperbola GNU (and future #BSD) with #mandoc
and maybe by promoting groff they are pretty much safe
even to typeset formulae into diferent formats, you can
pretty much use Groff for Math paper drafts and then
use TexLive for standard, rigid academic layouts.

gnuplot License | Software Package Data Exchange (SPDX)

#Linux became #slopware. The best #GNU/Linux distro still depends on #AI
slop. #Hurd is not ready yet. #Hyperbola #BSD needs support in order to
fully properly switch from a #GNU/Linux OS to a GNU? #BSD one.

Depressing times, but maybe enlightening ones too. Linux since 1996 was
just a corporate workauround to drive commercial Unix tools on it.

In the end #IBM won. My tip for the #GN[U] develoepers for #GUIX:
consider Unix dead. Don't try to resurrect it. It's finished, rotten,
bloated, propietarized, outdated. Ironically the licensed GPLv2 kernel
with less blobs than Linux came from AT&T and it's Unix 2.0.
They already did Unix better than Unix itself. Either you focus and enhance
#Hurd, or lose against the blobslop OS. Make an outstanding FS on Hurd.
Give more rights to the user than the shitty layers of #dbus/#polkit and the
like to fix the lack of proper namespaces on Linux.
I said 'proper', Plan9/9front it's light years ahead.
Don't get just 'well enough' with X11/wayland. You can do better stuff.
Ditto with Emacs, forget coreutils, reimplement everything you
can with jitted Elisp and Elisp code replacing these outdated Unix
tools. Eshell can be a better shell. The 9front folks already did it
with 'rc' simplyfing Unxi shells to the extreme.

Another reason to pick #hyperbola #BSD : #Torvalds' policy on #AI written
code for #Linux

https://github.com/torvalds/linux/blob/master/Documentation/process/coding-assistants.rst

I'll hardly miss that half #GPLv2 released kernel (full of dirty blobs
in the official release, not the #linux-libre one) with tools
and direction managed by #IBM in order to create their weird #AIX clone.

IDK about #GNU #Hurd, but is not ready. OFC you are free to rebase #hyperbola
with that kernel.

linux/Documentation/process/coding-assistants.rst at master · torvalds/linux

Linux kernel source tree. Contribute to torvalds/linux development by creating an account on GitHub.

GitHub

If #AI keeps corrupting almost every software project, I'll migrate
fully to #hyperbola #GNU (soon #BSD, but keeping the license
compatible with the GPLv2) and toss every modern crap.

9front too because the system it's small enough
and you need AI for that simple C and design
you should drop CS/IT altogether and go
back to High School or maybe Elementary school.
The system it's made to be understood
for humans, not to be lost with 100 millions
of code for a browser, a kernel and
some really complex desktop providing
userland techonologies as it happen with
current OSes.

I will just watch ripped DVD's under a netbook
with old releases of mpv/mplayer and pretty much
OPUS/mp3 and ogg libre media, even libre comics
under CBZ and stuff from Gutenberg.
No needs for DRM's, no need for LLM
slopware. I will just use Eforth+Subleq/muxleq,
rc+awk. And s9 for Lisp to try
'exotic' computing environments.
UXN maybe for 'retrogaming'.

Protocols? Plain IRC, NNTP, Jabber, Email and few more.
No need to vibecode crap there, you fools,
the standards are almost set in stone.
Use LibreSSL and that's it.

Keep backup from old systems, you will need them.
Create replicas of Hyperbola's mirrors.

Subleq+Eforth looks insane *but* it's a granted
environment under PD *and* from now you can
use it on almost every system where an ANSI C
compiler exists. It's almost useless for
anything modern but again it might even
work as a modern bootstrap tool if things
needs to be 'restarted' if a huge #LLM
cross-suiing between companies happens.

https://github.com/howej/muxleq

https://hyperbola.info

The versions are old enough to be trusted
to be #AI free. And the repo it's small
enough to be handled with modern
external media with TB sized disks.

Again, this is not madness, but a way to
be inmune to future cross-company suing
because of copyrighted code.

And not just code. For sure META will be damned
because of the Copyright steal.

@tusharhero I haven't look.
The GNU project has higher standards on
that. Guix by default won't accept any LLM
code I think, the stated that at guix-devel
mail list. But sadly as a daily machine I
find it slow and not as stable.

I'm still trying Trisquel to replace
Guix in my main machine, and keeping Guix
as a secondary PM to get updated
and free stuff without sacrificing
stability. It's good to have updated Icecat,
tons of libre games and policies to avoid
nonlibre code. If some LLM based software
has some issues, the version might be pinned
and the project could be even forked.

I'm pretty sure that Trisquel
might commit changes from Ubuntu unless they
patch it because of software ethics, such
as the SystemD age crap.

If things stay as they are, I'l roll #Hyperbola
GNU in that machine, compile an updated Libre kernel by hand
and, again, reinstall Guix as the 2nd PM even
if they #hyperbola people don't like it.

But real life works require software outside
of Hyperbola, trade degrees here make .Net/OpenHDK
almost mandatory.

Because of that I would just spawn
a Guix container, install all the crap inside
but *separate* from my profile, $HOME and the like,
and be perfectly safe with a Lumina desktop
and tainted stuff (even if it's FLOSS)
kept in a 'vault' outside of my normal
computing.

Another reason to hate Redhat/SystemD/Freedesktop:

https://github.com/flatpak/xdg-desktop-portal/pull/1922

#gnu #linux #flatpak #xdg #freedesktop

And that will spread over libre distros like #trisquel sadly.

#guix might be the only sane #gnu alternative among #hyperbola being rebased into a #BSD (nonGNU) but with a GPL license.

And maybe #9front, but 9front users will just spawn a cpu(4) auth outside of California and create a nightmare namespace hell for lawyers and troll lobbies such as #meta

Draft: Add parental controls to the Accounts portal by davidedmundson · Pull Request #1922 · flatpak/xdg-desktop-portal

Applications need to filter content to match the age rating of the user. The rating restrictions tend to be location and domain specific without a common ground for where these groupings should be....

GitHub

Watch Python Draw Hyperbola Curves

Square root math creates hyperbola branches square root calculations generate perfect hyperbola branches that curve Challenge accepted

#python #pythontricks #codingtips #programmingtutorial #hyperbola #mathcurves #conicsections #pythonmathart #viralpython #pythonshorts #curvedrawing

https://www.youtube.com/watch?v=8YJSQWoD8zM

Watch Python Draw Hyperbola Curves #pythonshorts

YouTube
@[email protected]

Dipende da cosa intendi per "libera".

Un tempo ti avrei detto Debian senza indugi, ed a tutt'oggi dal punto di vista legale, tutti i software contenuti nei repository free sono distribuite con licenze libere.
Il problema di #Debian è appunto politico (e dunque tecnico): le scelte del progetto, pur attenendosi alle #DFSG si preoccupano sempre meno della effettiva libertà degli utenti, anteponendovi altri valori.

Comunque, se ci limitiamo ad una libertà minima, formale, ci sono le distro elencate dalla #FSF https://www.gnu.org/distros/free-distros.en.html

Di queste, mi sembra promettente #Hyperbola, soprattutto per l'impegno a sostituire il kernel #Linux con un hard fork di #OpenBSD https://www.hyperbola.info/

Il problema comunque è l'effettiva esercitabilità di tutte e quattro le libertà, a fronte di una complessità insostenibile dei software distribuiti.

Avere #Firefox sotto MPL o #Chromium sotto MIT non garantisce davvero a tutti la libertà di studiarne completamente il codice: solo chi può investire mesi di studio (tipicamente pagati da qualche azienda) può veramente esercitare la libertà di studio. E ogni libertà esercitabile esclusivamente da un'élite smette di essere libertà e diventa, di fatto, privilegio.

Purtroppo Linux stesso (il kernel) ha una complessità esorbitante, misurabile in milioni di righe di codice, anche escludendo i blob non liberi.

Ne consegue che, ad oggi, chi vuole poter esercitare davvero le 4 libertà deve cercarle altrove.

I #BSD sono un po' migliori da questo punto di vista: ricordo una settimana passata a leggere il kernel di #NetBSD anni fa senza grandi problemi. Ancora meglio da questo punto di vista sono i #plan9 come #9front che in un paio di settimane può essere studiato da capo a piedi.

Il problema fondamentale però è più profondo di quanto non suggeriscano queste possibili soluzioni.

Quando #Stallman concepì il software libero, il maggior limite alle libertà degli utilizzatori (che erano anche programmatori) era di natura legale. Il #copyleft o le #DFSG erano strumenti legali (il primo) o comunitari (il secondo) progettati per evitare questo limite.

Oggi a limitare la libertà degli utenti troviamo invece vincoli tecnici (#SaaS, complessità del software) e culturali (standard intenzionalmente complicati da implementare ed una sostanziale separazione fra la figura professionale del programmatore ed il ruolo di utente).

Questi vincoli sono superabili con leggi ed investimenti in educazione e istruzione, ma difficilmente tali azioni possono essere proposte da politici ignoranti eletti cittadini mantenuti nell'ignoranza cibernetica.

D'altro canto produrre stack alternativi è difficile, non foss'altro che per l'assenza di risorse e coordinamento (nonché spesso di una visione architetturale coerente con l'obiettivo politico di massimizzare la libertà).

@[email protected] @[email protected]
List of Free GNU/Linux Distributions - GNU Project - Free Software Foundation

#Inverse functions also exist using the #hyperbola in the role of the #circle. In these cases in #trigonometry the #arcfunctions are included below (possibly innocent errors about imaginary numbers):

Hype for the Future 7A: Mathematics Edition

The hyperbolic functions of the sine and of the cosine are identified as follows:

sinh(x)=(e^x-e^(-x))/2

cosh(x)=(e^x+e^(-x))/2

The remaining hyperbolic trigonometric functions can be constructed with the hyperbolic versions of the appropriate functions. While tan(x)=sin(x)/cos(x), tanh(x)=sinh(x)/cosh(x) and can be constructed as (e^x-e^(-x))/(e^x+e^(-x)).