Gregory's Server
|
"Sorry I can't stream any video right now, we haven't got the bandwidth" I often think about the implementation-size/quality-of-life ratio when working, and I tend to be very accepting of my own UX shortcomings as to just work around them instead of fixing them. The other day I spent about 10 bytes to add these little buffer zones(cursor goes gray) between the various modules of Left's interface, and that's probably been one of the best bang for buck.  We hiked a mountain a few days ago, somewhere half-way up, we came across a massive bee/wasp nest built right into the ground of the forest, and we couldn't easily contour it, anyways, I got stung, and after a few days it still hurts :(  @neauoire I was on a run yesterday and something flew into my face and stung me. Pretty sure it was a wasp. Anyhow, beneath my right eye is still swollen so I'm wearing sunglasses indoors so as not to perturb people 😅  @neauoire Peace was never an option. 🤷♂️ It makes me laugh that these massive tech companies build these websites now that don't even bother printing a message on <noscript> it's like fuck you, a white page, it's all we can afford for you.  @neauoire It's better than adding an insult into <noscript>... Google Maps infamously said "When you eliminate the JavaScript whatever remains must be a blank page." Wrote a little slide rule desktop toy in 2600 bytes, it's not very accurate but it'll allow me to demonstrate how the logarithmic scales work in front of an audience. The concept we need is called "observational equivalence."  Anchored just a few minutes walk away from a fresh water lake hidden between two mountains, it has been a while since we could last properly wash ourselves. It's nice. Let's say you're implementing a map function to run in parallel, how do you decide how many threads to spawn. There seems to be a pretty heavy cost to spawning a thread per cell when the list is very long, instead of having each thread process a few cells each. What's this topic called? What do I need to search for best practices?
Show previous comments
 @neauoire If you have a perfectly async. system you never want more threads than cores. There simply is no point. Unfortunately the world is sync. even processors are sync. with a clock. Now if you really want to think about something that will tickle your brain try to imagine an async. processor with no clock. It has been done apparently but I still fail to imagine it. Started putting together the slides for Strange Loop, in one of the slides I'll need to do a bit of drawing, so I spent time revamping Adelie's interface mode, added speaker's notes in the console, added a timer because the conferences always forget to start the stage timer, and drawing a label that displays the upcoming slide's name. From time to time, someone will ask for a feature or spec change in Orca, and in about 99% of the times, it's a very setup specific way to handle some sort of message, that is best handled by creating a new custom operator. I love it when I get to be there at the moment when the person requesting the feature realizes they can add any operator watever that they need on top of the specs. "I CAN DO THAT!?" The chibicc compiler is growing so fast :') I will keep on declining any request to joining the instance where the person refers to themself as a refugee. "It disappears because we don’t design it, don’t build it, we only post into prepared forms." Next time someone sends me a link that takes more than a minute to open in Netsurf with our 5kb/s internet, I'll just answer TB;DR Too Bloated, Didn't Read.
Show previous comments
@neauoire this reminds me of an initiative I have tried to start multiple times in my professional roles Twice a year, for one week each, every engineer’s workstation is artificially limited to dialup-like speeds so they all have to feel the pain of their bloated web pages. It never really caught on sadly.  @neauoire Ow, that's somehow even /worse/ than our 15KB/s we get after running out of data on our phone. I've collected some notes on my little experiments into #uxn programming language stuff. Like program validation, graph reduction, structural editing, and all sorts of *cough* rather very dry topics, into something that should be approachable to anyone with a basic understanding of assembly languages. I've been wasting way too much time on this uxntal type checking thing, it's starting to feel like one of those programs that I make to save time, but in the end I just spend an unrecoverable amount of time fine-tuning. But despite that, it has one unexpected upside, which is that when it gets hard to predict the state stack, I've begun using defined labels's arity like breakpoints through the body of routines to validate that the stack state, at that moment in the program, is as predicted. I was already writing these comments from time to time to describe the transformations across different lines of a program, now they're actually part of the verification of the function. It's neat! I think I might write a thing about this once it's a bit less clunky. |
@neauoire This is such a beautiful photo Devine!!! Hope you get back all rested :)
@neauoire With what machine do you take your pictures? Is it a digital one? Is it something like a phone?