FoxQuirk 🦊 @quirk@computerfairi.es

imagine wanting to go from Shannon, Ireland to Toulouse, France but instead you are going to Belgium (a fate worse than death)

https://globe.adsbexchange.com/?icao=3964f2

watching a YouTube short about poaching large quantities of eggs, and the chef says that he steams them at 63 degrees [Celsius] for 45 minutes.

top comment:

For those wondering, 63° is 7π/20 radians.

interesting, Safari now outright refuses to load websites that use TLS 1.1, TLS 1.0, or RSA <= 1024

with no way to bypass

My daughter is collecting data on people's pineapple on pizza preferences, if you would be so kind as to answer, this will help to move us past this impasse, even if there is some bandwagoning involved

#pizza #pineappleonpizza

As a programmer slash reverse engineer with two decades of experience, I have seen many things. But today I saw something new. A full SaaS application with its entire backend interface built as … a single <canvas> tag.

That’s right, everything — user/permission management, uploaded file browser, system configuration screens — everything is painstakingly drawn programmatically as individual shapes and texts. Thank fuck I’m here as a user and not a developer.

#ProgrammingHorror

Scrub, Daddy!

If you try to tell people on tumblr it makes no sense they will get mad at you defending their "system."

But really what matters is that the people they know and like are there, and they have figured out how to interact with those people. That more logical systems exist doesn't matter. That systems that aren't "corporate walled gardens" doesn't matter. It's the people. And the people are why I keep going back from time to time.

The people are ALSO why I'm mostly happy here.

Tumblr's interface and commenting structure makes no sense.

You can:

comment on a post.
reblog it and add something to it. This creates a new copy of the post on your blog ... so if the original author makes edits they do not show up.
add hash tags, which users use for comments too for some reason
like a post (this does nothing)

It really is a nonsensical mess, but people have adapted to it and love it.

There is a lesson here. Online communities are made of people. People will adapt.

Today in “computers are cursed”:

I’m using pdf2cairo to generate a PNG preview of a PDF file.
PNG preview looks fine.
That PDF+PNG pair gets uploaded to a remote asset management system.
Remote system takes the PNG and generates its own previews from that.
These previews contain artifacts/garbled pixels. This behaviour is consistent, the same PNG always generates the same artifacts, and only in areas that are transparent.

Wanna guess what I had to do to fix this?

1/2

preferred rectangle data model:

wait you've all been taking ESTROGEN? this whole time i thought you'd been transitioning with the aid of the guy from "waiting for godot"

We made a comic about a detective.

A job I did a long time ago ⚡

@carrideen On a lark, a while back, I tried to use an AI to help with lesson planning. It gave me a list of 10 poems to use as companions to Fahrenheit 451. 6 of the 10 poems don't exist and 2 of the last 4 were not theme-based at all; I have no idea why it offered up those - listed on the same page somewhere maybe? And the thing is, I can spot this, but my students can't. I don't even know how to feel about this. All I can really do is caution my students about it, and gently "call them out" on AI content when I spot it. And by that I mean, query them verbally to ascertain if they wrote it, and if they didn't, explain our department's position and have them re-write those parts. It's a tool some people will use, and if they are, they need to do so appropriately.

In this case, no, the student was misremembering the name of a poet he read last semester. But, unable to identify a poem on that topic by Phillis Wheatley Peters, Google just made one up.

I've had problems all month trying to put the publication dates of texts next to the titles on the syllabus--quickly searching for a bunch of dates in a row, a lot of them are off by a few decades--just small enough that if you're not really thinking hard about it, might seem plausible.

@carrideen We’ve just spent a few decades teaching students that you should trust backed up information (= stuff you verify with internet search) above a single person’s opinion, as your classmate or neighbor can be wrong.

And now when teachers say ”that’s completely false” about something a kid read from a sentence generator, it can rightfully seem like the teachers are the ones mistaken. There is the valid-seeming search result, after all.