@jbqueru@fosstodon.org
@jbqueru@fosstodon.org avatar

jbqueru

@jbqueru@fosstodon.org

I'm JBQ / Djaybee, Husband, Immigrant, Veteran, Highly Sensitive Person #HSP. He/Him. I write about tech and other things. I'm fluent in French and English.

I like: #skiing, #hiking, #biking, #games, #photography, visual #astronomy. #PixelArt, #painting, #knitting, #weaving, #crochet. #bead weaving, #CrossStitch and #BlackWork embroidery.

I am in the year-long process of moving from Spokane, WA, USA to Preveza, Greece.

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

jbqueru, to random
@jbqueru@fosstodon.org avatar

Write the first version as if you're going to throw it away. You will, anyway.

(Paraphrased from Fred Brooks).

mwichary, to random
@mwichary@mastodon.online avatar

What would you consider as the most recognizable bitmap fonts in tech history?

I’m imagining stuff like:

  • the arcade/Atari font
  • Chicago (Mac, then iPod)
  • VCR/video equipment fonts
  • Minecraft font
  • IBM PC fonts (MDA, VGA, stuff like that)
  • perhaps System font from Windows 3.x
  • Commodore 64, just because of the sheer popularity of the machine

What am I missing?

image/png
image/png
image/png

jbqueru,
@jbqueru@fosstodon.org avatar

@chrisphin @mwichary Yup, I was about to say, anyone who has used a the Amstrad font enough (whether on a CPC or on a PCW) can probably recognize it very quickly.

timbray, to random
@timbray@cosocial.ca avatar

EU looking at introducing liability for providers of software products: https://www.europarl.europa.eu/legislative-train/theme-a-europe-fit-for-the-digital-age/file-new-product-liability-directive

Note explicit carve-out for F/OSS software. I wonder how this affects managed-OSS-as-a-service products like AWS’s RDS and OpenSearch?

jbqueru,
@jbqueru@fosstodon.org avatar

@timbray I haven't looked at that one (yet), but I'm hoping that they align the F/OSS exceptions with those in the CRA (Cyber Resilience Act).

jbqueru, to random
@jbqueru@fosstodon.org avatar

I'll be a lot more impressed by AI-based code-generating tools when I can tell them something like "port the game Solaris from the Atari 2600 to the Amstrad CPC" and they output something that works.

(I'm currently studying the theoretical possibility of such a port. It might be feasible, but it's not a simple task).

jbqueru,
@jbqueru@fosstodon.org avatar

@mmu_man @pulkomandy I'm not sure whether I'd be honored (for getting help from someone with experience) or ashamed (for showing my ugly ugly code).

We're on the same forum where I'm pretty sure I can get help, and my code is on my github page pretty much as I write it (oooh, exposure!)

For now, I'm OK, I'm feeling my way through Z80 coding that's still fairly new for me, and I've got something that feels about right.

scottfgray, to random
@scottfgray@dice.camp avatar

“This machine kills AI.”

jbqueru,
@jbqueru@fosstodon.org avatar

@jeroen94704 @idiran @charadon @scottfgray I asked it once to generate code for a specific retro console ("draw a rectangle"), and its first 2 attempts used programming languages that aren't available for that console.

(1/2)

jbqueru,
@jbqueru@fosstodon.org avatar

@jeroen94704 @idiran @charadon @scottfgray Eventually, it used the right language, though the output didn't compile, and it got tons of things wrong to the point where the code wasn't anywhere close to drawing a rectangle.

It seemed to be able to integrate several library calls together in a way that somewhat matched my prompts, but, when it came to writing actual code that actually does things (as opposed to calling other people's code) it got stuck.

(2/2)

jbqueru,
@jbqueru@fosstodon.org avatar

@jeroen94704 @idiran @charadon @scottfgray I recently asked it to comment some assembly code I had just written but had not commented yet. Commenting uncommented assembly code is an extremely common task in reverse engineering. It had no clue what the code actually did, commented instructions almost one-by-one and even got some of them wrong (!!!), such that it was useless at best, harmful at worst.

jbqueru,
@jbqueru@fosstodon.org avatar

@squishymage42 @jeroen94704 @idiran @charadon @scottfgray That I very much agree is a case that current state of the art is probably able to handle reasonably well.

jbqueru, to random
@jbqueru@fosstodon.org avatar

Oh joy, Ubuntu upgrade introduced some crazy lag in the terminal. Very much not fun.

jbqueru,
@jbqueru@fosstodon.org avatar

Switched display from NVIDIA to Intel GPU, problem solved. NVIDIA still available for compute. And now I can use Wayland.

poundquerydotinfo, to random
@poundquerydotinfo@virctuary.com avatar

LB: They're finally discontinuing the Z80 series :(

jbqueru,
@jbqueru@fosstodon.org avatar

@poundquerydotinfo Noooooooooo!

Sadly, yes, it seems so.

Time to go write some more Z80. I've been stalled for over a week, getting stuck over a performance issue, I need to get myself unblocked.

jbqueru, (edited ) to random
@jbqueru@fosstodon.org avatar

Apple (Apple II): we'll intentionally match our pixel frequency to the NTSC color burst frequency so that it creates more artifacts.

Atari (VCS/2600, 800/5200, 7800): we'll intentionally match our pixel frequency to the NTSC color burst frequency so that it creates fewer artifacts.

foone, to random
@foone@digipres.club avatar

People are always telling me I'm "basic" but they won't clarify if they mean like qbasic or visual basic or one of the 8bit micro basics?

jbqueru, (edited )
@jbqueru@fosstodon.org avatar

@foone GFA? STOS? AMOS? 1000D?

jbqueru, to random
@jbqueru@fosstodon.org avatar

No matter how I look at it, I can't get myself to call the TI 99/4A a computer in its default configuration, because 256 bytes of CPU-accessible RAM doesn't make a computer in my personal opinion. As shipped, it's a game console with a keyboard, with the same graphics chip and substantially the same sound chip as the ColecoVision and the SG-1000.

In turn, that makes it one of the early consoles to have a 16-bit CPU, possibly the first one.

coreyspowell, to space
@coreyspowell@mastodon.social avatar

The astronauts aboard the ISS were amazed by the April 8 solar eclipse, just like the rest of us, and went crazy snapping photos from orbit.

They got some amazing views of the Moon's shadow sweeping across southern Canada & Maine, seen from 420 km (261 miles) above: https://www.flickr.com/photos/nasa2explore/with/53645140833

The Moon's shadow, or umbra, is pictured covering portions of the Canadian provinces of Quebec and New Brunswick and the American state of Maine in this photograph from the International Space Station as it soared into the solar eclipse from 261 miles above.
The Moon's shadow, or umbra, is pictured covering portions of the Canadian provinces of Quebec and New Brunswick and the American state of Maine in this photograph from the International Space Station as it soared into the solar eclipse from 261 miles above.

jbqueru,
@jbqueru@fosstodon.org avatar

@coreyspowell On the other side (literally), a lunar eclipse is an opportunity to see approximately how big the earth would be as seen from the moon (by completing the curvature of the earth's shadow), at least within half a degree out of about 2 degrees.

cassidy, to random
@cassidy@blaede.family avatar

I have to take a short weekend trip to Montana and after considering different transit options, I decided to rent a car and drive; it's a bit of a road trip, but I found great rental deals and was excited to rent an electric car to try it out as we're considering buying one.

lol, Montana is huge and has hardly any charging stations on the route 🙃 According to one of the EV charging maps there are single chargers in towns about 150 miles apart, which seems dicey to me?

jbqueru,
@jbqueru@fosstodon.org avatar

@cassidy I drove EVs for years, that gives me tidbits of experience.

There's a constant tension between having chargers scattered everywhere and having large groups of chargers.

The former is convenient for local drivers and reduces risks of general failures. The latter makes maintenance easier and increases the change of availability at any site.

(1/2)

jbqueru,
@jbqueru@fosstodon.org avatar

@cassidy My gut feeling:

Every town with at least one store should have at least one group of 2 chargers. So should every store with a dedicated parking lot.

Along major travel routes, there should be charging sites every 40 miles, with at least 10 chargers per site or more according to demand. (40 miles means that a car with 200 miles of highway range can start at 80%, drive past 2 sites, and reach the 3rd site with 20% left).

(2/2)

usul, to random French
@usul@piaille.fr avatar

Quand l’hyper-threading devient contre-productif
https://www.palabritudes.net/2024/03/28/quand-hyper-threading-contre-productif.html

#performance #tests

jbqueru,
@jbqueru@fosstodon.org avatar

@usul Mon CPU est annoncé à 32 threads et 5.8 GHz.

Mais 16 threads sont en SMT, et seuls 2 des 8 cœurs impliqués tournent à 5.8.

Les 16 autres threads sont sur des cœurs plus simples, à 4.3 GHz, sans SMT mais qui partagent le cache L2 par groupes de 4.

Sur les 32 threads, je peux en avoir 24 sans SMT, 12 sans partage de L2, 16 sur les cœurs puissants uniquement, 8 sur ces cœurs sans SMT, 16 sur les cœurs simples uniquement, 4 sur ces cœurs sans partage de L2, etc.

Tout ça partage le L3.

jbqueru,
@jbqueru@fosstodon.org avatar

@usul Linux privilégie les 2 premiers threads sur les 2 cœurs les plus rapides.

Ce qui est marrant, c'est que j'utilise soit des applications peu parallèles, avec rarement plus de 2 threads actifs, ou des applications très parallèles qui saturent les 32 threads.

jbqueru,
@jbqueru@fosstodon.org avatar

@usul Surtout du rendu. PovRay, Blender (mais le GPU est plus rapide!).

jbqueru, (edited ) to random
@jbqueru@fosstodon.org avatar

As I'm unpacking after moving, I'm still missing one computer I'm looking for... but my BeBox is out!

jbqueru,
@jbqueru@fosstodon.org avatar

@brab I tested it before moving. It seems to have booted, but no video output. I have a spare video card, I'll need to dig deeper.

jbqueru, (edited ) to random
@jbqueru@fosstodon.org avatar

If you think that my recent test with ChatGPT and assembly is irrelevant for being unrealistic, the actual task of disassembling code (for CPU or VM) to try to understand what it does is common. That's e.g. some of the work that's happening to understand the harmful xz payload, which is provided as a binary.

My experiment wasn't even a worst-case scenario: The code implemented a single feature, was not obfuscated, had meaningful labels, and had newlines to help group the instructions.

jbqueru, (edited ) to random
@jbqueru@fosstodon.org avatar

I just gave ChatGPT another chance at helping me write some code. I pasted some uncommented chunk of assembly (not trick code), and asked it to comment that code.

The most it was able to guess "intelligently" is that the code had to do with graphics. Beyond that, though, it was quite useless. It tried to comment instructions one by one with their literal meaning, and even got that wrong. It also couldn't group sequences of instructions into logical groups.

(1/4)

jbqueru,
@jbqueru@fosstodon.org avatar

@troed writing is easier for such systems as long as they have something to copy from.

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