@ljrk@todon.eu
@ljrk@todon.eu avatar

ljrk

@ljrk@todon.eu

gecos/CN: Leonard/Janis R. König
SAN: ElleJay (lj), Janis, Leo

content:

  • Commentary on world and German politics
  • Teaching and related politics
  • Rants about bad typography (& occasional praise)
  • Linux/LaTeX/GNOME/whatever-editor-I-currently-try-using
  • IT Security & Reversing
  • Climate change and anti-car policies
  • The Gender™

Opinions my own.

pfp: me, red lipstick, red foxy cat ears, brown hair tucked into a blue-grey shawl, silver feather earrings.
banner: photo of various hacker & related stickers

JMT: https://justmytoots.com/@ljrk@todon.eu

This profile is from a federated server and may be incomplete. Browse more on the original instance.

catsalad, (edited ) to chrome

Google announced that starting in June 2024, ad blockers such as uBlock Origin will be disabled in Chrome 127 and later with the rollout of Manifest V3 ().

The new manifest will prevent using custom filters and stops on demand updates of blocklist. Only authorized updates to browser extension will be allowed in the future, which mean an automatic win for Google in their battle to stop YouTube .

is deceitful and threatening to your privacy, and now is a good time to switch to (@mozilla) and/or (@torproject) if you haven't done so already!

EFF (@eff) on Google’s Manifest V3:

⚠️⁠https://www.eff.org/deeplinks/2021/12/chrome-users-beware-manifest-v3-deceitful-and-threatening
⚠️⁠https://www.eff.org/deeplinks/2021/12/googles-manifest-v3-still-hurts-privacy-security-innovation

Chrome Manifest V3 Transition Timeline (2023-11-16)

🚩⁠https://developer.chrome.com/blog/resuming-the-transition-to-mv3/

EDIT for clarification: MV3 in Chrome will still allow some ad blocking extensions, but will severely limit their blocking ability and even restricts pre-set filters to 50 MAX.

ljrk,
@ljrk@todon.eu avatar

@catsalad @mozilla @torproject @eff The second paragraph is of particular importance since that also explains why AdGuard is supporting it (see Google's announcement) and the "wider ad blocking community" (i.e., all the bullshit that's not the one true blocker aka uBO) is okay with it. They sell ad blockers and lists as a product. They're okay with not blocking everything (like YT), especially if noone else can do it anymore because the competition (uBO) is sidelined.

This push is not against ad blockers actually, it's against effective ad blockers, particularly uBO.

TechConnectify, (edited ) to random
@TechConnectify@mas.to avatar

NOTE TO LITERALLY THE ENTIRE INTERNET:

No thought is ever complete. There will always be holes to fill!

RESIST THE URGE TO FILL THOSE HOLES

If someone has said something which you largely agree with, especially if what they said aligns with your understanding of and goals towards an issue, then the LAST thing they need from you is someone playing devil's advocate. Nitpicking is annoying at best and, in a world where people are looking to drive wedges between allies, dangerous at worst.

1/2

ljrk,
@ljrk@todon.eu avatar

@TechConnectify Yes, never should you make someone feel bad for mostly being on your side.

But then again, I don't think "filling in holes" is such a bad thing. In fact, that's exactly what I want for me from the Fedi (am I weird? probably). I love to get nitpicked because it usually means I'm about to learn about someone's special interest.

And I love learning about that, just as I love sharing my special interests or weird oddly specific research. I don't want to attribute everything to autism and across that spectrum we likely have people who feel differently about that as well.

But I can't help to categorize this into autistic "love language" which is often about sharing information, esp. oddly specific one that's "filling in holes". And I/we often get perceived as wanting to "be right", but that's not the case. We want the thing to be correct, we don't have a personal stake in it: We want to be corrected as well, when we're wrong (even if only in details).

As with unsolicited advice, I love to get that. I've always thought that this must be solved differently, we need some way to tell people "hey, I like nitpicks and odd details or weird advice" vs. "pls, I just want to be let alone and focus on the main parts". Both are valid, but censoring one feels very neuronormative :/

RickiTarr, to random
@RickiTarr@beige.party avatar

I wonder how things are going on Twitter...

Portal X @Portal Elon Musk, the Tesla billionaire and founder of X has sadly passed away this morning at the age of 52. 74 Readers added context they thought people might want to know Elon Musk is not the founder of X, formerly known as Twitter. The original founders of Twitter are Jack Dorsey, Noah Glass, Biz Stone, en.wikipedia.org/wiki/Twitter, ... britannica.com/topic/Twitter LU NIRRT [
Ron DeSantis & i @RonDeSantis “Success is not final, failure is not fatal: it is the courage to continue that counts.” - Winston Churchill I T " @ Not shown on X The quote is actually from a Budweiser ad in 1938
Andy Boenau @Boenau It's often safest for a cyclist to go through a red light. Here's an example of safely keeping momentum. Bicycles and motor vehicles should never be treated as equals, so "but red means stop!" isni't a useful reaction. Readers added context they thought people might want to know As per the Highway Code rule 69 cyclists "MUST obey all traffic signs and traffic light signals." With further guidance in rule 71

ljrk,
@ljrk@todon.eu avatar

@RickiTarr O.o nobody correcting Musk not being dead (except dead inside) and the last context is... not really addressing the point (cyclist safety rather than legality)?

Makura, to random
@Makura@pl.mintyriver.xyz avatar

lmao

RT: https://mastodon.social/users/dannotdaniel/statuses/112261156490807195

ljrk,
@ljrk@todon.eu avatar

@Makura Yah (although tbf I see more hate on Masto than vice versa, but different bubbles). I'm also making a point of saying "The Fedi" personally, despite being on Masto.

girlonthenet, to random
@girlonthenet@mastodon.social avatar

Because it's , every woman is allowed to propose one (1) new law. It must be a silly, low-stakes one which has little impact on society other than to solve one of your pet peeves.

Mine is: 'make it illegal to illustrate articles about female pleasure with photos of someone fingering a piece of fruit.'

WOMEN: what's yours?

ljrk,
@ljrk@todon.eu avatar

@girlonthenet OH: Make it illegal to randomly declare something "women's" or "men's", e.g., drinks, shaving cream, razors, shampoo, ...

hateaid, to random German
@hateaid@troet.cafe avatar

Historisches Urteil für alle Betroffenen: OLG bestätigt, dass Meta massenhafte Falschzitate selbst finden und löschen muss. Ein wichtiges Zeichen gegen systematische Desinformation. Mehr zu unserem Erfolg im Grundsatzprozess mit Renate Kuenast: https://hateaid.org/historisches-kuenast-gegen-meta-urteil-haelt-stand/?mtm_campaign=grundsatzprozess-kuenast-berufungsurteil-pm&mtm_kwd=mastodon

ljrk,
@ljrk@todon.eu avatar

@hateaid
Ich weiß nicht ob ich die Freude über das Urteil teile. Ich verstehe sehr den Bedarf nach einer Lösung, auf der anderen Seite haben wir bereits schon diverse Massenlöschungen mit Massenscans wg. vermeintlichen Verstoß gegen Kopierschutz (oft betroffen: unabhängige Künstler*innen und politische Inhalte), sexuellen Inhalten (viele queere/feministische Inhalte werden geblockt) und "automatische Scans" die KiPo erkennen sollen.

Wir kämpfen seit Jahren gegen dieses massenhafte Scannen von Inhalten wie auch bei der Chatkontrolle angedacht. Und auch wenn das hier für einen guten Grund gemacht wird, das Missbrauchspotenzial ist riesig.

Ich denke nicht, dass eine reine technische Lösung (oder Vorgaben) der richtig weg ist. Wir sollten stattdessen mit unseren technischen Plattformen unsere realen sozialen Netzwerke nachbilden. Dadurch greifen unsere natürlichen "Filter": Hetzer müssen erstmal in die soziale Bubble reinkommen und diese ist sehr sorgfältig in der Auswahl. Solche föderierten Systeme (Mastodon, PeerTube, ...) gibt es auch bereits schon und ihre Moderation ist Lichtjahre besser als die von zentralen Plattformen – ohne das Verbrennen von riesigen Mengen an Energie.

thephd, to random
@thephd@pony.social avatar

It's time.

RAII in C, and why nobody's getting it right with the increasingly "simple" juggling that keeps getting tossed to me like table scraps for a dog.

The Pasture | Why Not Just Do Simple C++ RAII in C? | https://thephd.dev/just-put-raii-in-c-bro-please-bro-just-one-more-destructor-bro-cmon-im-good-for-it

ljrk,
@ljrk@todon.eu avatar

@thephd As someone who usually prefers C over C++ for various kinds of reasons (some of which you described very accurately as C's "merits and goals") this article speaks to me in so many ways. Making C a stripped down version of C++ doesn't do justice to either language.

Personally, if I had the energy and spoons, I'd probably drive a different agenda for C altogether though, namely effectively deprecating it and viewing every proposal through the lense of "does it inhibit switching away from C/make it more complicated to use C in a dual-language project or as glue code?"

IMHO we should add a lot of small annotations (where possible) that "simply" (cough) allow programmers to tweak the language to become more strict. nonnull (through [static 1]) is a good start. Basically allow the dev to tighten the rules and knobs of C s.t. the compiler can warn or error more or do better analysis. The whole point of it would be to help maintain old code bases as well as aid migration and discourage new projects. But that's my rather personal outlook 0:-)

whitequark, to random
@whitequark@mastodon.social avatar

please tell me the most obscure joke you know

(feel free to explain or not explain it, depending on what you find more amusing to think of me reading it)

ljrk,
@ljrk@todon.eu avatar

@whitequark Q: What's an anagram of Banach-Tarsky?

A: Banach-Tarsky Banach-Tarsky.

ljrk, to random
@ljrk@todon.eu avatar

Okay, the American Dialect Society is on fire:

> effective altruism: movement ostensibly to benefit humanity, used as an excuse for spending other people’s money

https://americandialect.org/wp-content/uploads/2024/01/2023-Word-of-the-Year-PRESS-RELEASE.pdf

The best definition of

dgar, to random
@dgar@aus.social avatar

You can tell a computer geek from an influencer by the way they pronounce “bios”.

ljrk,
@ljrk@todon.eu avatar

@jpaskaruk @dgar And PNG ("ping", as specified in RFC!)

ljrk, to random
@ljrk@todon.eu avatar

Ich poste selten Wirtschafts Woche, aber ich finde toll, dass sie sich zu dieser Schlagzeile trauen:

"Wochenend Fahrverbote: Ist das ernsthaft Ihr Niveau Herr ?"

https://www.wiwo.de/unternehmen/auto/wochenend-fahrverbote-ist-das-ernsthaft-ihr-niveau-herr-wissing/29754576.html

ljrk, to passkeys
@ljrk@todon.eu avatar

Wow, the comments on my article on in the German / has shown me a lot of misconceptions people have:

  • No, you don't need to synchronize Passkeys
  • nor do you need to use Google/MS/Apple
  • nor is storing an encrypted binary blob a big danger
  • Passkeys aren't just autofilled : they use challenge auth, not shared secrets!
  • 's aren't better because they're a real . Actually they suck against .
  • A secure enclave can still be used, but it's mostly used for decrypting the keychain, not storing it
  • You can still use 's, either with discoverable creds (uses 1 slot each) or non-discoverable creds (1 slot for all Passkeys)

Generally, I think the term 2FA is misleading. Not all 2FA is created equal. One could even argue that Passkeys are "less" 2FA than Password+TOTP -- and yet, it's more secure in most attacks because it can't be phished.

A lot of people seem to think that the more annoying and difficult to use a technology is, the more secure it is. We have the same problem with passwords and their complexity. We humans suck at guessing how secure something is through intuition.

Article:
https://www.heise.de/meinung/Kommentar-Passkeys-sind-toll-fuers-Internet-und-schwierig-in-Unternehmen-9543202.html

ljrk, to random
@ljrk@todon.eu avatar

Currently internally debating whether I'll attend . Since my endurance dumped: Before I was able to run 25km without prior training, now I'm gasping after 5. I cannot do any strength/core training without carefully tracking my heart beats.

I do recommend everyone who hasn't been at any Congress to visit if they can, it's an incredible experience. But I had my share of Chaos events this year already and the a) non-existence of any precautions b) the statement of unwillingness to go beyond what's required by law (since when do we care about that?), combined with c) the IMHO disgusting phrasing on the announcement taking the cherry on the top, only solidified by the comments of orga related people really makes me question going there.

I do understand that it's difficult to mandate masks and I can accept a decision based on this rationale if communicated that way. One could formulate the social goal of reducing infections despite the difficulties. But instead everything is shifted to the individual's responsibility, a weird "libertarian" idea for which I only have utter contempt. There are people worse off than me who need to rely on us behaving properly.

Masks don't hurt. But they're effective.

ljrk, to random
@ljrk@todon.eu avatar

Pure chaotic energy:
https://github.com/MrGlockenspiel/activate-linux

ljrk, to llm
@ljrk@todon.eu avatar

Just discovered (thanks @elora) @baldur 's post on comparing •s with psychics and it's scarily accurate. Now reading through the backlog of the other articles written and holy hell they're incredibly good.

Take this:
https://www.baldurbjarnason.com/2023/ai-research-again/

Quote:

» […] autocompletes need to be substantially better than an average coder to avoid having a detrimental effect on overall code quality. «

Yes, yes and yes! This is sooo important. We really need more people in tech who do not disregard research of other fields, e.g., psychology and how humans interact with tech. We're slowly getting there in UX research, but AI is the newest wart of horrible naïve tech without a good use case.

I wrote about my own annoyances of here:
https://ljrk.codeberg.page/ethical-ai.html
Although I primarily focus on its impact on decision making & morals, I also cover the problem of "conserving" the current state of society/programming, effectively halting/slowing down progress.

Blickwinkel, to random German
@Blickwinkel@digitalcourage.social avatar

Guter Beitrag von Casey Kreer @kc
- Rettet uns die KI?

Über die Zukunft der digitalen Inklusion

Spätestens seit Ende 2022 sind generative KI-Systeme wie ChatGPT und Midjourney in aller Munde, und sie werden dabei nicht selten auch als Game-Changer für die digitale Barrierefreiheit postuliert. Doch wo stehen wir eigentlich gerade wirklich, was können diese Systeme bereits jetzt für uns tun, und was bringt uns die Zukunft? Es ist höchste Zeit für einen unverfälschten „Reality Check“ und einen authentischen Blick in den Alltag von Menschen mit Behinderung.
https://youtube.com/watch?v=BmRU_HkPwXg&si=6roHHi66CClWL4F_

ljrk,
@ljrk@todon.eu avatar

@Blickwinkel @kc Es ist – wie mit vielen Technologien – einfach so, dass man sie sicherlich für Inklusion nutzen könnte, dafür aber kein Wille ist. Autos hätten damals sehr früh für mehr Inklusion sorgen können, haben dann aber jahrzehntelang Rollstuhlfahrende eher stärker Möglichkeiten genommen durch die Stadt zu navigieren.

Züge hätten vorher ein ähnliches Potential gehabt, stattdessen hat man sich nicht darum gekümmert die bekannte Lücke zwischen Bahnsteig und Bahn zu schließen.

Und KI könnte man natürlich stärker nutzen um Dinge darauf zu entwickeln die die Welt inklusiver machen. Aber für eine inklusivere Welt muss sich erstmal die Gesellschaft entwickeln, dann kommt die Technik von allein. Viel zu gerne möchten wir uns mit Technik frei kaufen, dass wir jetzt – mit KI – ja gar nix mehr tun müssten.

Der größte Humbug, im Gegenteil, durch diese Vernachlässigkeit wird's schlimmer.

/rant

ljrk, to random
@ljrk@todon.eu avatar

Recently learned about "Oppositional Defiant Disorder" and honestly these:

  • Often argues with authority figures or, for children and adolescents, with adults
  • Often actively defies or refuses to comply with requests from authority figures or with rules

... just sound like being anti-authoritarian... and not a disorder (perhaps neurodivergence). I may be biased because that sounds like a description of a relevant part of my persona but what the fuck I don't care for your authority lol

ljrk, to Matrix
@ljrk@todon.eu avatar

GNOME will soon be able to do session verification! hype

https://gitlab.gnome.org/World/fractal/-/merge_requests/1629

ljrk, to random
@ljrk@todon.eu avatar

The difference b/w imperative programming and declarative:

Imperative

  1. Cook 6 tomatoes
  2. Put them into ice water and wait until cooled
  3. Peel and dice them
  4. Finely chop 1-2 onions
  5. Heat up some olive oil, and fry the onions until slightly transparernt
  6. Add the tomatoes, some sugar, thyme, salt and pepper
  7. Bring to boil and cook it until the desired thickness is reached

You now have a tomato sauce.

Declarative

A tomato sauce is made by cooking 6 peeled and diced tomatoes with some added sugar, thyme, salt and pepper as well as 1-2 chopped onions slightly fried in olive oil.

While imperative programming is a step-by-step instruction what to do, declarative programming is a description what is. Think tutorial/recipe vs. encyclopedia entry. Thus, a declarative approach completely ignores time and sequencing: In theory, if you already have store-bought peeled tomatoes, you can just use them. If they're already diced, you can do that as well. If they come pre-seasoned with salt and pepper, that's okay as well.

In terms of programming: If an expression has already been evaluated, it can easily and trivially be re-used.

ljrk,
@ljrk@todon.eu avatar

It's much harder to reason whether those store bought peeled tomatoes would work for the imperative recipe because a) it conflicts with the instructions and b) maybe they were peeled differently. The declarative version makes it implicit that actually we do not care about how they were peeled. If someone at some points invents genetically modified tomatoes that grow pre-peeled, they'd be fine as well.

The imperative approach restricts the solution of the problem to precisely what was implemented. If we would want to prepare some steps simultaneously, we'd need to do that explicitly. With declarative code, the parallelization is simply an implicit possibility.

A prime example is make(1) and Makefiles: You state what dependencies need to be satisfied for any rule and make figures out what can be done in parallel. Just pass -j N with N being the number of parallel jobs.

Of course, you need to be careful anytime declarative code is interacting with the imperative world. So with make: The commands executed are stateful and may not always be able to run in parallel.

Similarly, the interpreter/runtime must have its own scheduling logic: Preparing food from the declarative description above requires much more thought on the cook's side than the recipe.

ljrk, to random
@ljrk@todon.eu avatar

After having spent a /lot/ of time teaching, I'm quite convinced that most of the "I'm bad at X" things are faults of the teacher, not the student.

I've seen people who "could never do math" rock their 1.0 in exams, understand and do quite tricky cryptography or do some crazy functional programming leveraging higher algebra.

All that was required was a change of teacher, proper teaching method and some time (mostly to overcome the "I can't do this" thinking).

Of course some people are better or worse at some things. But too often this is an excuse of bad teachers not to improve and effectively victim blaming.

(There's also the issue that modern teaching institutions have the wrong goals and too little time to actually enable good teaching on a wide scale.)

ljrk, to random
@ljrk@todon.eu avatar

ist, wenn die Kommentare unter dem Artikel tlw. mehr Ahnung haben als der eigentliche Artikel. oof.

ljrk, to random
@ljrk@todon.eu avatar

Ratet Mal wessen Bahnverbindung zur von der Bahn gecancelt wurde und wer keinerlei Benachrichtung dazu bekommen hat? Hm?

Danke für Nix liebe DB. Was ein Scheißverein.

BrodieOnLinux, to random
@BrodieOnLinux@linuxrocks.online avatar

This is the Linus we all know and love https://lore.kernel.org/lkml/CAHk-=wi5YPwWA8f5RAf_Hi8iL0NhGJeL6MN6UFWwRMY8L6UDvQ@mail.gmail.com/

ljrk,
@ljrk@todon.eu avatar

@yrlf @BrodieOnLinux I don't think he's right w.r.t. the ergonomics though and the final goal though:

  1. It's not about simply turning on a flag and generating thousands of warnings which drown "real issues" (his argument in one comment). Kees said that he would go through that heroic effort of flagging and changing the types where needed. By introducing a new type though, it's finally possible to tick those "off". Afterwards, it's not many errors at once anymore.
  2. If the type would be called u32_wrap, no-one would "just cast to that" without thinking about this actually needing wrap, it's right in the type.
  3. Actually Kees' effort could even be used to find some kind of pattern that would help detect which wraps are intended and which aren't. We'd have a quite big sample size then.
  4. He's advocating for fixing symptoms rather than root causes. It's seldom successful to enumerate all the symptoms (or even just close to all) of a root cause w/o building a perfect detector for every dangerous wrap around – a very hard task indeed.

Since such a detector is not known, it's good to err on the side of caution, as otherwise attackers use such an overzealous analyzer and will exploit every of such instance that's actually a vulnerability.

ljrk, to random
@ljrk@todon.eu avatar

Okay, I thought about getting a humorous necklace poking fun at Christian crosses but upside-down crosses are too much associated Satanism. And while I appreciate the effort and support Satan, I'm not much of a believer at any rate.

So I thought of a pole dancing jesus (which gets better once I hopefully start doing pole myself) but instead of a pole, it's his cross?

ljrk,
@ljrk@todon.eu avatar

@vman I'm a horrible person using ChatGPT for this...

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • megavids
  • kavyap
  • DreamBathrooms
  • GTA5RPClips
  • magazineikmin
  • InstantRegret
  • cubers
  • thenastyranch
  • Youngstown
  • rosin
  • slotface
  • osvaldo12
  • ngwrru68w68
  • ethstaker
  • JUstTest
  • everett
  • Durango
  • normalnudes
  • Leos
  • mdbf
  • khanakhh
  • tester
  • modclub
  • cisconetworking
  • anitta
  • tacticalgear
  • provamag3
  • lostlight
  • All magazines
    •