Nostr notes by SOMETHING WONDERFUL HAS HAPPENED TO YOUR leah :oh_no:

when you realise you are dealing with actual monsters: ...

2026-10-28T02:07:51Z

when you realise you are dealing with actual monsters:

https://codeberg.org/growler/k/src/branch/master/m.c

does anyone have a suggestion of a retro-oriented instance, like ...

2026-10-09T17:46:18Z

does anyone have a suggestion of a retro-oriented instance, like oldbytes, but with a strict anti-LLM policy? doesn't have to be Mastodon, although i really like markdown editing and four-digit character limits

it feels like i'm having to block half the local timeline right now, and that's annoying

or, dare i say it, The Philosophy For People Who Think ...

2026-09-14T14:01:29Z

In reply to nevent1q…m002
_________________________

or, dare i say it, The Philosophy For People Who Think They're Too Smart For Objectivism

re LB: Effective Altruism, or Why Wrong Is Really Right, or We ...

2026-09-14T14:00:02Z

re LB: Effective Altruism, or Why Wrong Is Really Right, or We Can Prove That Grass Is Purple

you know, if you would ever post alt text it'd make it a lot ...

2026-09-08T18:23:29Z

In reply to nevent1q…52zs
_________________________

you know, if you would ever post alt text it'd make it a lot easier (at least for me) to figure out what you were imaging about

i can't find the toot that pointed me at this article any ...

2026-09-01T15:56:45Z

i can't find the toot that pointed me at this article any more, so i'm reposting it myself:

https://brennan.day/normalized-fascism-in-open-source-12-million-given-to-dhh/

so after beind nerdsniped re the origin of apropos: it appears to ...

2026-08-31T19:27:41Z

so after beind nerdsniped re the origin of apropos: it appears to have arrived in Unix with Bill Joy's man implementation in 2BSD, in 1979. but Joy was certainly aware of Lisp - indeed, Berkeley gave the world Franz Lisp, written in C for Unix, in 1980 - and i'm almost entirely sure that Lisp Machine Lisp already had apropos when the first Chinual was published in 1976

Nostr event nevent1qqszr5sqz4uh59ntk2cmwl4fcg9dym4fdwwvwxm3u7zfj867a49w2xqzyz4xlvayrt0sghq3dy6e5a69m7s9twx7clum67mdc5tqulhrwxjh2dzlmeh

2026-08-30T15:50:21Z

another claude botherer, another block

oh, i see. now everyone thinks it's acceptable to break a ...

2026-08-30T15:44:02Z

oh, i see. now *everyone* thinks it's acceptable to break a previously working system with an "update". regression testing isn't "for the rest of us", right?

i'm failing to read an article in lwn about 128-bit page ...

2026-08-27T20:02:30Z

i'm failing to read an article in lwn about 128-bit page tables for ARM, and i think mainly my brain is just refusing to accept that i will ever need to know anything about this

(every particle in the universe can be addressed individually with 270 bits, if you're wondering about the maximum practical address space)

apparently parentheses are an unexpectedly Goldilocks-y aspect of ...

2026-08-27T19:48:56Z

In reply to nevent1q…s8u5
_________________________

apparently parentheses are an unexpectedly Goldilocks-y aspect of programming

shows someone Lisp "ugh, does it have to have that many ...

2026-08-27T19:47:32Z

**shows someone Lisp**

"ugh, does it have to have that many parentheses?"

**shows them Forth **

"no! not like that!"

someone asking, utterly perplexed, how to do a right shift when ...

2026-08-24T05:25:09Z

someone asking, utterly perplexed, how to do a right shift when the only instruction they have is subleq reminds why i have no patience for hyperminimal instruction sets

Nostr event nevent1qqsqcvmhjrv3yw4vauzxqpgemdh85843gtpmm8ksn7px3adxqnxpc6gzyz4xlvayrt0sghq3dy6e5a69m7s9twx7clum67mdc5tqulhrwxjh20aj6gz

2026-08-20T16:09:27Z

is git.sr.ht down for everyone or just me?

is there a unix tool to do arithmetic on dates? as ...

2026-08-20T10:05:32Z

is there a unix tool to do arithmetic on dates? as in<code>$TOOL '100 days' -after 'may 9th' </code>

or something?

love too wait half an hour for rust to build something, only to ...

2026-08-13T22:27:46Z

love too wait half an hour for rust to build something, only to see it fail almost exactly halfway through with "you don't have any libclangs installed" when i have *two* libclangs installed because nobody can agree on a fucking *version*

i watched Millennium before i watched Soylent Green, and now i ...

2026-08-09T21:22:37Z

In reply to nevent1q…sugp
_________________________

i watched Millennium before i watched Soylent Green, and now i likely never will because of the spoiler :-(

at least i found out about To Serve Man before the last eason of Buffy

with regard to Rust specifically, i suspect there's a much ...

2026-08-07T00:12:38Z

In reply to nevent1q…cjz5
_________________________

with regard to Rust specifically, i suspect there's a much smaller, simpler language waiting to be discovered within Rust. it might be better to pursue that language, rather than trying to be a better C++

re LB: sadly at this point, i think the only way to get AI-free ...

2026-08-07T00:11:33Z

re LB: sadly at this point, i think the only way to get AI-free anything is via hard forks

fuckin techbros, basically /rant

2026-08-01T16:18:37Z

In reply to nevent1q…kah3
_________________________

fuckin techbros, basically

/rant

it's devastating to me that i've watched Linux go from ...

2026-08-01T16:17:39Z

In reply to nevent1q…hayq
_________________________

it's devastating to me that i've watched Linux go from the "don't worry, we'll keep it going forever!" OS to the "ffs are you still using *that*? here's a nickel, buy yourself a better computer" OS

indeed, there's already one example, directly relevant to me, ...

2026-08-01T16:15:59Z

In reply to nevent1q…axg5
_________________________

indeed, there's already one example, directly relevant to me, in which i know something works perfectly in NetBSD but doesn't work at all in Linux - my HP T510. an X server fires up just fine under NetBSD. damn thing doesn't work with anything under Linux, and hasn't for *years*

"oh but you won't be able to do everything in NetBSD" ...

2026-08-01T16:14:06Z

"oh but you won't be able to do everything in NetBSD" look, i can't do everything in Linux any more, so how exactly would i be worse off?

TIL EFI shell 1.1 included both a text editor and a hex editor. ...

2026-07-26T03:57:40Z

TIL EFI shell 1.1 included both a text editor and a hex editor. kind of surprised it doesn't include the equivalent of DEBUG too

when even OSDev's approach to calling UEFI is "don't ...

2026-07-26T03:39:17Z

when even OSDev's approach to calling UEFI is "don't worry your pretty little heads about it, the C compiler and headers have you covered", you know something has gone fundamentally wrong with computers

> All of the AI projects we have observed as a team are ...

2026-07-18T13:31:13Z

> All of the AI projects we have observed as a team are failing. Every single one – we have seen 0% success in a year and a half, not only amongst projects we have been asked to participate in2, but even within projects that we have observed in passing while doing totally unrelated work.
> -- > [> https://>; ludic.mataroa.blog/blog/ai-man> ia-is-eviscerating-global-decision-making/](https://ludic.mataroa.blog/blog/ai-mania-is-eviscerating-global-decision-making/ )

sometimes the plural of anecdote *is* data

> I don’t understand this complaint. The maintainer didn’t ...

2026-07-14T15:28:06Z

> I don’t understand this complaint. The maintainer didn’t write this software for you

i never cease to be amazed at how many ways software developers can find to reject any and every notion of responsibility to their community.

if the author really didn't write it for anyone but themselves, why did they publish? if the author is also the maintainer - the only instance in which your statement might even be testable! - then why do they fix bugs they have not encountered?

if you maintain software for a community, you are implicitly accepting the responsibility to do so properly. if you don't want to do that, vacate the position for someone who will. better that software be thrown over the wall than maintained badly.

also, it turned out to be yet another dialect of Lisp

2026-07-11T15:20:36Z

In reply to nevent1q…s2ve
_________________________

also, it turned out to be yet another dialect of Lisp

remember when Firefox didn't randomly fall over after a few ...

2026-07-10T15:05:04Z

remember when Firefox *didn't* randomly fall over after a few hours for no apparent reason whatsoever?

surprised to discover that Voidlinux still packages XForms! so i ...

2026-07-07T00:46:23Z

surprised to discover that Voidlinux still packages XForms!

so i immediately tried to compile LyX 1.4.5.1 with it, that being (afaik) the last version that supported XForms; but unfortunately, it failed because strcmp and strncmp are no longer members of std

the wonderful thing about open source: if you want to compile one thing from 2007, you damn well better have *everything* come from 2007

the past is another operating system; they do things incompatibly there

i remember you posting the results on here

2026-07-07T00:17:03Z

In reply to nevent1q…xxzz
_________________________

i remember you posting the results on here

on the other hand, if the Athena widgets had been released with a ...

2026-07-07T00:15:08Z

In reply to nevent1q…u3wr
_________________________

on the other hand, if the Athena widgets had been released with a default .Xdefauls that made them actually attractive, we maybe wouldn't have poured *quite* so much time into trying to come up with alternatives

of course, the single reason Motif didn't take off in the ...

2026-07-06T23:47:15Z

of course, the single reason Motif didn't take off in the open source community is that it wasn't. it is *now*, far too late to make a difference; but back when, if you wanted to write open source graphical apps, you basically had to use XView or one of a bunch of variations on Athena.

XForms had the same problem. although it went LGPL a decade earlier than Motif. so did Qt at first, until it realised it could pioneer the dual-licensing model - but by then GTK already existed, and the X11 world had been sent to hell forever :-( had XForms been open from the start, we'd probably never have ended up with the whole KDE vs GNOME mess - we'd all be using XFCE instead

> Even the vanilla X11R4 “xclock” utility consumes 656K to ...

2026-07-06T23:33:52Z

> Even the vanilla X11R4 “xclock” utility consumes 656K to run. And X’s memory usage is increasing
> -- UNIX-HATERS' Handbook, p.123

oh... you sweet summer children...

> Some say the learning curve is steep, but you only have to ...

2026-07-05T16:40:14Z

> Some say the learning curve is steep, but you only have to climb it once.

said someone who never had to use tar or find

i've successfully compiled stuff that used autotools from ...

2026-07-05T04:54:16Z

In reply to nevent1q…vcna
_________________________

i've successfully compiled stuff that used autotools from decades ago. and yet, CMake won't work with something released in August 2024.

i can only conclude that CMake isn't intended for use in the same way that autotools is; it's only intended for short-term use within closed development projects.

that's fine, but they need to say so, and explicitly counsel against source releases that rely on CMake.

this is for software released *less than two years ...

2026-07-05T04:47:29Z

this is for software released *less than two years ago*:<code>$ cmake . CMake Error at CMakeLists.txt:11 (cmake_minimum_required): Compatibility with CMake < 3.5 has been removed from CMake. Update the VERSION argument <min> value. Or, use the <min>...<max> syntax to tell CMake that the project requires at least <min> but has been updated to work with policies introduced by <max> or earlier. Or, add -DCMAKE_POLICY_VERSION_MINIMUM=3.5 to try configuring anyway. -- Configuring incomplete, errors occurred! </code>

fuck you ***so very much***, CMake, and fuck everyone who still uses you

<code>unknown font: verdana </code>oh ffs. ...

2026-07-05T04:41:06Z

<code>unknown font: verdana </code>oh ffs. ok...<code>$ find /mnt/omni -iname verdana.ttf ./fontback/msttcore/verdana.ttf $ sudo cp -R /mnt/omni/fontback /usr/share/fonts $ fc-list verdana /usr/share/fonts/fontback/msttcore/verdana.ttf: Verdana:style=Regular,Normal,obyčejné,Standard,Κανονικά,Normaali,Normál,Normale,Standaard,Normalny,Обычный,Normálne,Navadno,Arrunta […] $ ./Self -s kitchensink.snap […] unknown font: verdana $ linux go home, you're drunk _ </code>

i do so love the fact that it's perfectly normal to compile a ...

2026-07-03T01:08:24Z

i do so love the fact that it's perfectly normal to compile a program with Xft enabled, then start it by specifying program -fn FONT, then have it immediately fail with a "can't set font" message.

it is 2026 and this is **bullshit**.

now that brings back some memories... ...

2026-06-29T22:06:54Z

In reply to nevent1q…ksk9
_________________________

now that brings back some memories...

https://www.youtube.com/watch?v=557Npb85Emo

re LB: so basically any commercial VPN you might choose has an ...

2026-06-28T16:11:02Z

re LB: so basically any commercial VPN you might choose has an even chance of being a Trojan horse for domestic or foreign intelligence... and if it isn't - or even if it is - then it's either already a washing powder brand, or soon will be one when they realise they can't make money at it and sell up.

thinking about it, the worst aspect of this is that that CEO must ...

2026-06-28T00:21:58Z

thinking about it, the worst aspect of this is that that CEO must have started Mullvad *with the specific intention* of using the profits he made from it to fund something expressly counter to its stated goals. Mullvad has basically been one big Trojan horse for *its entire existence*.

that takes cynicism to another level.

well... that's that for RISC OS Pi then. the ancient copy of ...

2026-06-17T18:17:23Z

well... that's that for RISC OS Pi then.

the ancient copy of RISC OS i've been using with my Pi 1 is creaky, and for some reason stuck with a 16-bit GPU framebuffer, but it works perfectly with my Barco E-3620 and the custom monitor definition i made for it. (from some of the strangeness, i assume there was some display scaling going on.)

but i just tried the latest RISC OS, v5,30 - and i dropped the same MDF onto the display manager... but - nothing. i can only select 640x480. RISC OS seems to have completely lost the ability to do custom modes. i mean, i do at least actually get 256 greys in greyscale modes, rather than just 32 - but... if i can't have them at 1536x2048, it rather defeats the purpose.

anyone who knows RISC OS and can suggest a solution, please let me know!

quick question: do TVs that recognise PAL and NTSC automatically ...

2026-06-01T18:39:12Z

quick question: do TVs that recognise PAL and NTSC automatically also recognise hybrids like PAL-M and PAL-60?

yes, by all means, if you want to underline the point i was ...

2026-05-27T12:06:04Z

yes, by all means, if you want to underline the point i was making about OpenBSD contributors being a *special* kind of asshat, just send me a link to the FAQ with no comment, and when i ask why, reply only "RTFAQ". you will *definitely* encourage me to switch to your favourite operating system with that attitude.

🤦‍♀️

**plonk** ( ( ( ☎️ ) ) )

is it possible that i've uncovered the only OS fandom objectively worse than the hordes of reply guys and 'splainers that constitute Linux's...?

sidenote: whoever decided to call it Wayland instead of Y needs ...

2026-05-25T01:06:43Z

In reply to nevent1q…9cq3
_________________________

sidenote: whoever decided to call it Wayland instead of Y needs to re-evaluate their life choices... but then i'm still firmly of the opinion that the successor to C should have been called P

all this talk of Tripos is reminding me of another OS that relied ...

2026-05-25T01:05:48Z

all this talk of Tripos is reminding me of another OS that relied on a microkernel and superoptimised message passing and really needs to be updated for modern systems.

no, not QNX, if you wnt there. its grandfather, Thoth... whose offspring included the system on which the predecessor of the X windowing system was developed.

oh for FUCK'S SAKE ...

2026-05-12T18:34:03Z

oh for FUCK'S SAKE

https://codeberg.org/small-hack/open-slopware

the eZ80 is much less wtf-y than the 65816 in many ways (so few ...

2026-05-04T16:01:08Z

the eZ80 is much less wtf-y than the 65816 in many ways (so few transistors, so many bad decisions!), but whoever decided that it was OK to have the only practical way to access the high byte of a 24-bit address be storing it into memory and loading from n+2 should have to subtitle their CV "i have no design taste"...

i mean, at least<code> ld (temp),hl ld a,(temp+2) or push hl ld hl,2 add hl,sp ld a,(hl) or ld (ix-1),hl ld a,(ix+1) </code>

only take 12, 11 or 10 cycles respectively; but still...

in general, "someone else will do that icky stuff for ...

2026-05-02T02:12:24Z

in general, "someone else will do that icky stuff for you" is not a sustainable programming practice.

Raymond Chen was writing about Windows, not generally. it's ...

2026-05-02T02:05:03Z

In reply to nevent1q…hxmg
_________________________

Raymond Chen was writing about Windows, not generally. it's not hard to imagine a system which expects memory to persist between processes - or one in which that expectation is made use of.

what a bunch of arseholes ...

2026-05-02T01:26:50Z

what a bunch of arseholes

https://serverfault.com/questions/110219/which-os-distributions-have-64-bit-kernel-and-32-bit-userspace

it's a perfectly reasonable question, yet look at *literally every single one of them* refusing to answer it.

that's linux "helpfulness", in a nutshell

hang on... so you're seriously telling me that the option on ...

2026-05-02T00:39:55Z

In reply to nevent1q…qyvk
_________________________

hang on... so you're seriously telling me that the option on which everyone seems to have settled is that malloc() always *pretends* to work, but that the *entire system* might grind to a halt at some future point as the operating system struggles so hard to release free pages to satisfy that promise that it eventually thrashes all code executed anywhere in the system through a single page...?

are you kidding?

are they high?!

no doubt such folk would prefer a system in which malloc() blocks ...

2026-05-02T00:36:27Z

In reply to nevent1q…qk2r
_________________________

no doubt such folk would prefer a system in which malloc() blocks on failure (because obviously a shortage of memory is a temporary problem) rather than returning null... wait. what do you mean, they don't like that option either?!

certainly, being forced to code for such systems would put the ...

2026-05-02T00:33:06Z

In reply to nevent1q…e974
_________________________

certainly, being forced to code for such systems would put the all-the-world's-a-POSIX folk who insist that you don't have to check malloc()'s result because what can you do if it failed? in something of a tight spot...

i wonder how much modern software ultimately relies on the fact ...

2026-05-02T00:30:13Z

i wonder how much modern software ultimately relies on the fact that modern operating systems will reclaim all their resources?

because i read recently about a comparatively widely used system (was is Tripos?) where that wasn't true - programs were expected to return all their resources to the OS before quitting, because the OS wouldn't do it for them

one thing i still want to see is a tiny rust compiler, akin to ...

2026-04-27T13:29:15Z

one thing i still want to see is a tiny rust compiler, akin to tcc. something that can do all the checking and compile all the code, but inthe most direct way possible, rather than going all in on optimisation.

perhaps we could call it trust?

every so often i wonder what happened to machine-oriented high ...

2026-04-24T19:18:59Z

every so often i wonder what happened to machine-oriented high level languages - things like PL/360, PL/516, or Action!

and then i remember. C happened. an entire class of languages, killed at a stroke! (though oddly enough, BCPL didn't have nearly the same apocalyptic effect on MOHLLs for word-addressed machines that C did for those on byte-addressed machines)

more and more, i want to hear about software that FOSS ...

2026-04-18T23:04:07Z

more and more, i want to hear about software that FOSS evangelists sneer at because its licence contains clauses like "no fascism" or "no warmongering"

sigh i mean, i get why every new kernel released by anyone ...

2026-04-18T10:18:46Z

**sigh** i mean, i *get* why every new kernel released by anyone is POSIX compliant and can compile all their favourite programs. recreating the corpus of software everyone relies on these days, especially for internet connectivity, is a literally terrifying prospect. and of course computers are useless without internet connectivity, aren't they? ;-)

but it's killed the field of OS research - especially hobby OS research - stone dead. nobody's doing anything that departs from the "kernel in C implementing POSIX" model, even though *literally every other seam* is all but untapped.

froggey's Mezzano is an honourable exception, and one i really need to install on a respectable machine - it's just, i also need to learn Common Lisp and emacs, and that's a *lot*, all at once! but if it can work well on one of the (um, ten) Wyse 3040s i ended up with, then i might well have a permanent little ersatz Lisp machine in the corner

so has anyone made a linux distro built on mlibc yet? if so: ...

2026-04-18T10:07:54Z

so has anyone made a linux distro built on mlibc yet? if so: which one? comments, opinions, etc?

never mind "free, open source first, community-owned" - ...

2026-04-16T23:01:59Z

never mind "free, open source first, community-owned" - point me at a git hosting service that bans AI contributions to any of its repos outright

the exp(x)-ln(y) finding is a fun result, but it doesn't ...

2026-04-15T16:10:44Z

the *exp(x)-ln(y)* finding is a fun result, but it doesn't appear to be of the same order of practicality as "you can build everything out of NAND gates". partly because exp and ln are complicated functions to calculate in themselves, especially when y is negative! so they don't really simplify anything in the way being able to fab an entire chip of gates the same way does - which is itself sort of accidental; as i understand it in NMOS fabrications an inverter can be expanded into an n-input NAND gate (and in PMOS a NOR gate) by just adding n-1 extra bases to the input transistor - essentially making a compound serial sequence of transistors. (the other kind of gate is made by parallelising the input transistors)

fun fact (also TIL): in CMOS a NAND gate is made by stacking the input transistors of a PMOS NAND gate on top of those of an NMOS NAND gate! same for NOR gates. the effect of which is that you don't need the output amplifier that NMOS and PMOS gates do, which means the whole arrangement only consumes power while transistors are switching; the design benefit is that you can choose whichever gate is most advantageous, rather than preferring NAND or NOR in NMOS and PMOS respectively, because they take the same surface area

i knew NetBSD had declared LLM-regurgitated code verboten. i ...

2026-04-14T11:34:17Z

i knew NetBSD had declared LLM-regurgitated code verboten. i didn't know that edict was passed nearly two years ago - does this mean that NetBSD has dodged the worst of it?

according to https://en.wikipedia.org/wiki/List_of_POSIX_commands ...

2026-04-01T03:15:15Z

In reply to nevent1q…2r43
_________________________

according to https://en.wikipedia.org/wiki/List_of_POSIX_commands they all originate with AT&T - sort from v1, join from v7, paste from 32V and cut from System III

(but if you go to the cut and paste WP entries, it turns out they were originally BSD additions)

and why, exactly, do cut and sort have different options for ...

2026-04-01T00:59:03Z

and *why*, exactly, do cut and sort have different options for selecting fields...?

Nostr event nevent1qqs0mk4uzrp78wejmx7vz9cl9t77y0x9l9z5tyfef6pu46hgtmcwc4szyz4xlvayrt0sghq3dy6e5a69m7s9twx7clum67mdc5tqulhrwxjh239ul2u

2026-03-30T18:53:04Z

tell me how whimsical claude can be, get muted

how bad do you have to be at business to fail to make money ...

2026-03-27T07:28:50Z

In reply to nevent1q…7tpd
_________________________

how bad do you have to be at business to fail to make money selling porn?!

re LB: like i said a few days ago, "i don't do cults" ...

2026-03-24T19:21:05Z

re LB: like i said a few days ago, "i don't do cults" has become the single most apposite answer to anyone asking why you didn't wrote your project in Rust

the assumption embedded in this bit, in particular, annoys this piss out of me:

> it has gotten to a point where people have told me that i should… stop because “you are making the world worse by writing insecure code”

C code is *not* automatically insecure! in fact, anyone assuming that Rust is automagically preventing them from writing insecure code should frankly step away from their keyboard, turn off their computer, and go and find a hobby they can cut themselves on. then they might learn a thing or two about the complex interrelationship between tools and personal responsibility.

i can see why that accusation *would* wreck an autistic person. it's one of those "you aren't even making it to the level of being wrong" category errors. it's impossible to sensibly argue against, because it's nonsense! but how on earth does one effectively counter nonsense...?

i'm beginning to think that Smalltalk is literally the only ...

2026-03-12T08:21:16Z

i'm beginning to think that Smalltalk is literally the only language family that does booleans properly

https://lwn.net/Articles/1059177/

...ok, maybe Scheme and the Pascal family too, but none of them quite *run with it* like Smalltalk

like, i just tried to install tooi, which has failed 3 times. the ...

2026-03-11T19:26:54Z

In reply to nevent1q…rkuc
_________________________

like, i just tried to install tooi, which has failed 3 times. the first time, because it needed cffi. the second time, because it needed aiohttp. an dd the third time, because it somehow objected to the system aiohttp i just installed and attempted (and failed) to roll its own.

i don't know how to resolve that error, so my attempt to build tooi is at an end, because *life is too damn short* to deal with the *absolute bullshit* of the people who think a fucking *scripting language* is a suitable medium for app distribution.

if python is so "batteries included", why does building ...

2026-03-11T19:23:59Z

if python is so "batteries included", why does building literally *anything* that uses it always involve dragging in 37 different packages, at least one of which is guaranteed not to exist / compile?

Nostr event nevent1qqsz0fjzr3xg2yl3s2pclv9xrtgk9kh8f98ej74r5mnzkjweyqtvfjgzyz4xlvayrt0sghq3dy6e5a69m7s9twx7clum67mdc5tqulhrwxjh2neu7ez

2026-03-10T22:49:19Z

"this meeting could have been an IRC session"

remember back in the days of DOS, when it was just accepted that ...

2026-03-09T16:23:43Z

remember back in the days of DOS, when it was just accepted that some folk would still be using DOS 3.30 because they didn't have any compelling need to upgrade? or back when the general consensus was that if you were using Win95 OSR2 or Win98SE, Windows ME was, at best, not worth the bother?

could something like LZ4 or ZX0 compression strike a happier ...

2026-03-07T22:36:38Z

In reply to nevent1q…av4m
_________________________

could something like LZ4 or ZX0 compression strike a happier medium?

https://github.com/einar-saukas/ZX0
https://github.com/pararaum/lz4-6502

"but i'm bad at coding". yes. so is everyone else. ...

2026-03-07T20:42:28Z

"but i'm bad at coding". yes. so is everyone else. getting good at coding is the humbling process of realising how bad you are at it and figuring ways to get better. not by going "oh well a computer must be better than me at programming, i'll let it do it for me"

damn. i hate it when someone is asking about a situation i have ...

2026-03-07T20:40:11Z

damn. i hate it when someone is asking about a situation i have almost exactly faced, and which i am *almost* sure i remember how to deal with, but....not quite sure

like, i so want to be helpful but i even more don't want to be accidentally unhelpful... **bites tongue**

the chardet situation demonstrates that LLMs are literally ...

2026-03-06T08:46:09Z

the chardet situation demonstrates that LLMs are *literally* plagiarism machines. worse - they are plagiarism *launderers*. in no other reality would replacing all the code in chardet with a clean-room implementation and then claiming that this represents a "new version" be possible, let alone permissible - at the very least, it'd have to be called 'new-chardet' or something distinct. and then it'd have to fight for its place in the ecosystem as a better implementation of the old chardet.

instead, someone appears go have gone "gee, i wish i didn't have to obey the LGPL for this library" and gone looking for ways around it, whilst also stealing the valour of the existing code.

frankly the only acceptable response to this is for the Python maintainers to explicitly disallow any chardet greater than v6, *and* any future contributions from this maintainer. *pour encourager les autres.*

funny thing. in contrast to the veritable gush of people who ...

2026-03-05T16:11:05Z

funny thing. in contrast to the veritable *gush* of people who have suddenly adopted vibecoding and even though you'd think they'd know better they're basically all saying the same things about how great it is...

i can't help but think back to Chuck Moore's recent - and by his own description, miserable - experience of trying to get Microsoft's own LLM to figure out how to use the Microsoft API that stopped working the way colorForth expected it to.

one small, simple problem - and he couldn't cajole the LLM from the company that made it to help him solve it.

so i guess the lesson to derive is that LLMs are great if all you care about is spewing out code that passes the test suite. but actually *debugging* that code? not gonna happen. the LLMs can't do that.

*because stochastic parrots aren't capable of analysis*.

"Oh my god, was I Claude’s nanny?!" Erin Cikanek's ...

2026-03-04T02:39:10Z

"Oh my god, was I Claude’s nanny?!" Erin Cikanek's key realisation about having AI code for you...

https://erincikanek.com/the-rumors-of-my-demise-have-been-greatly-exaggerated/

i'm just going to say this: i remember when Linux could ...

2026-03-03T23:56:36Z

i'm just going to say this: i remember when Linux could cheerfully run a GUI and a web browser in 64MB RAM. for all people like to say "oh, but web browsers did less then", they had CSS, JavaScript and multimedia capabilities by the turn of the century (the timescale i'm talking about). they also had XSLT, Flash and Java, and they could read email and news, and connect to FTP servers (and i think gopher servers too). so much for "they do so much more now". no, they really don't - they just make a hell of a mountain out of a rather smaller molehill.

> One of my main software goals was to be able to run the ...

2026-03-01T14:03:48Z

> One of my main software goals was to be able to run the program “cool-retro-term” or CRT, which uses 3D shaders to simulate the look of a cathode ray tube with an LCD panel

um. not to yuck anyone's yum, and i do appreciate that CRTs are increasingly hard to find, and more importantly, that even running a Pi 5's VideoCore flat out and overclocked will come *nowhere near* the power consumption of a CRT...

but honestly, i don't get the appeal of emulating the worst visual aspects of CRTs - badly (remember, back then we all used to use inch-wide margins on our CRTs precisely *because* of the curvature at the edges!) - when LCDs are just *better*. in almost every respect - and the respects in which they aren't won't be solved by emulating a CRT with them

on the subject of "operating systems created by only one ...

2026-02-25T11:32:54Z

on the subject of "operating systems created by only one person", anyone else remember AtheOS?

https://atheos.metaproject.frl/

if nVIDIA charged money to up your frame rate, it'd be well ...

2026-02-23T03:36:36Z

In reply to nevent1q…ljwe
_________________________

if nVIDIA charged money to up your frame rate, it'd be well into the three figures

i mean, i'm sure the technical capacity exists - i'm sure some of their lower speed stuff is artificially slowed down higher speed stuff, and it'd only take a line of firmware to unlock it - but chip manufacturers have learned that never *ever* releasing that line of firmware, under *any* circumstances, is what keeps the business of selling of speed grades alive

they're in the GSoC '26

2026-02-21T20:33:34Z

In reply to nevent1q…8q7t
_________________________

they're in the GSoC '26

um, congratulations to Haiku OS for winning the devil's deal ...

2026-02-21T20:31:53Z

um, congratulations to Haiku OS for winning the devil's deal contest, i guess

what has really gone wrong with softwre, why bloat is running ...

2026-02-11T05:33:27Z

what has really gone wrong with softwre, why bloat is running away with itself?

*nobody programs real hardware any more*.

seriously, what is it about the words "but with ...

2026-02-10T13:51:07Z

seriously, what is it about the words "but with *computers*!" that makes people stop recognising ancient scams?

what if... home computers had come with something like the Canon ...

2026-01-24T21:12:22Z

what if... home computers had come with something like the Canon Cat OS installed? where basically the interface to the machine is a text editor, but if you select a chunk of text and press the RUN key, it drops into its implementation language (probably Forth, maybe BASIC) and runs it? and games can be loaded by compiling source code into memory on the fly, or by loading the text file, compiling the source code and then immediately overwriting it with object code and game data?

the SwyftCard was this, and was apparently written in Forth, although the language it exposed was Applesoft BASIC (presumably because that meant it didn't need Forth headers). hmm...

i think the falling out between me and Void Linux is imminent i ...

2026-01-24T00:35:10Z

i think the falling out between me and Void Linux is *imminent*

i can't update google-fonts because i don't have 2GB free on my 8GB root directory, even though /usr/share/fonts is cross-mounted on a disk with 270GB free

there is theoretically a voidlinux installation on there already, so i decided to try using xbps-install to update that

but no. it demands i update xbps first. so i try to do that, but there is a script that fails because of an exec format error. because that is an x86_64 system, and i am trying to update it from an i686 system, and apparently Nobody Has Ever Done That™.

of course, it would be a whole lot easier if google-fonts were split into its component packages, like qemu has been. but no. i don't know whether licensing prohibits or what, but if i want all the latest google fonts on your system (don't judge me), i have to somehow free up 25% of my root drive.

as with everything, what "the Left" (whatever tf that is) ...

2026-01-23T14:50:44Z

In reply to nevent1q…qmz5
_________________________

as with everything, what "the Left" (whatever tf that is) *does* hate is narrow-minded prescriptivism

Nostr event nevent1qqsv7v374hngfvx6889mqwt8tngjr7z09gyxs6umfjur6dfmqt0mxuqzyz4xlvayrt0sghq3dy6e5a69m7s9twx7clum67mdc5tqulhrwxjh2haujy6

2026-01-13T21:08:58Z

oh no, has OSNews stomped tech.lgbt off the web?

i think the biggest problem i have with GNOME is its tendency to ...

2026-01-11T01:08:25Z

i think the biggest problem i have with GNOME is its tendency to insist that the new shiny thing is The Only Possible Way To Do Things Now, with the singular focus of a toddler on a sugar high chasing a shiny ballloon

kind of puzzled as to why more people don't use the RISC OS ...

2026-01-10T15:11:37Z

kind of puzzled as to why more people don't use the RISC OS outline font format, given that it's a format designed to be used on integer-only processors with single-digit MIPS speeds and memory measured in kilobytes, for which the specifications have been available for decades...

at the very least, a portable reimplementation could have been interesting

domestic electricity costs about 3 times as much as that in the ...

2025-12-27T01:05:35Z

In reply to nevent1q…ascl
_________________________

domestic electricity costs about 3 times as much as that in the UK

i must confess, i find the stance that the Linux kernel is all ...

2025-12-25T03:13:15Z

i must confess, i find the stance that the Linux kernel is all about sharing the code back to upstream, and not remotely about ensuring that people can hack on the devices that they *own, having paid good money for*, somewhere between self-absorbed (it helps the Linux kernel to be adopted more widely, but does nothing at all to help people take meaningful ownership of their devices) and just plain pointless (the only thing you can actually *do* with the source code as supplied by a company that locks down its device's kernel is look for backdoors - you can't fix anything, which was the whole reason the GPL was created in the first place)

what if we started with the premise that sharing memory between ...

2025-12-22T23:35:58Z

what if we started with the premise that sharing memory between cores was the root of all of our security issues, and simply... stopped doing that?

one thing i really miss: the days when someone said "oh, ...

2025-12-22T22:02:56Z

one thing i really miss: the days when someone said "oh, i've got a $TYPE_OF_COMPUTER" and everyone knew exactly what the specification of that machine was and exactly what software it could or couldn't run

now? we're not *quite* in the realm of "no two computers have exactly the same specification", but we're a hell of a lot closer than we were. and yet every computer, no matter how dissimilar to anything else, is expected to run all the [same] software.

exactly. tbh i mind emulators that run on microcontrollers - even ...

2025-12-20T16:27:33Z

In reply to nevent1q…9ga3
_________________________

exactly. tbh i mind emulators that run on microcontrollers - even ones as powerful as the RP2350 or ESP32 - a whole lot less than i object to someone who just chucked a Cortex A series SoC running VICE on Ubuntu into a funky box and went "eh, job done"

programming a PIC24 is enough like retrocomputing anyway that ...

2025-12-20T16:24:19Z

programming a PIC24 is enough like retrocomputing anyway that anyone using one for anything interesting automatically gets a free pass