The Antikythera Mechanism Episode 11 – Inscribing the Back Plate – Part 1
97 points • 22 comments
@gillesjacobs
From 1/29/2018, 3:44:36 PM till now, @gillesjacobs has achieved 1073 Karma Points with the contribution count of 210.
Recent @gillesjacobs Activity
Sounds neat and I might try for the novelty. I am only skeptical of the regular user's need for and usability of rolling release OSs: This is not a volunteer project and not officially supported by Canonical, which means something is more likely to break (e.g., when Canonical changes their devel setup).
A rolling release system has to be really robust, and I get that that is the main difficulty faced by any continuous release system. But especially OSs should take great care in validating for breaking changes. No-one wants half their workday gone because their OS broke on a daily upgrade.
I bet Canonical is keeping a close eye on this project because rolling releases are probably on Ubuntu's product feature wishlist somewhere!
Retiring at 62 in a 440k USD house on a neotropical beach development feels like a pipedream to me and I am better off than most of my generation.
Yeah I rely on Docker. Anaconda gave a lot of issues too, but that is not my choice for python package mgmnt anyway. All issues are fixed with specifying Linux/amd64 as the build platform and only installing conda in containers. Installing conda locally is not something you should ever do in any case, it makes a mess of your home, shell profiles and environment variables.
Still having OOM 'Killed' errors in builds that I should investigate. Don't know for sure if there are memory limits in default Docker settings on Macos, could be it. In any case, I should respec cleaner builds anyway because right now it's a mess.
I used it a lot for university exam cramming.
Looks really cool and it's Python too! Interesting how all assets are string inserts unto the map [1]. Looks like a conceptually solid engine and great introduction to game dev.
1. https://github.com/lxgr-linux/scrap_engine/blob/master/scrap...
Thanks that's one everyday frustration down!
Exactly my experience too. It's amazing how people get invested in the brand and emotionally attached to the vendor locked walled garden. When you ask for something as simple as universal HiDPI scaling on Apple forums. A feature which many Linux distros have for 5 years now and Windows for even longer. I got passive-aggressive comments of how I should just buy the high-end Apple or LG displays and Apple docks which are literally 3x more expensive.
Any notebook with Windows and multiple Linux distros work with HiDPI scaling on any combination of my various docks, monitors, and cables.
Apple definitely does not "just work" in this case. They really "think different" as in different standards to ensure consumer lock-in in their walled garden.
'Just werks' is great marketing but not backed up by reality. I have the displeasure of being handed an M1 Mac for a new ML and NLP job.
Let's not even talk about the issues osx-arm64 architecture brings for DevOps.
Bluetooth: devices do not disconnect when the device ID put to sleep. If I don't disconnect my headphones before leaving my laptop, I can't connect to my phone. My guess this is done for some Apple Bluetooth features, anyone using not-Apple is SooL.
HiDPI: Not using an Apple monitor? Too bad no hdpi scaling for you.
Yeah the hardware quality and design is nice, software and OS is hot garbage. My Linux Mint ThinkPad had higher UX polish than Macos Monterey.
This sucks and it is surprising how badly they handled the situation. They should at least offer existing customers priority for contact renewals. That's customer retention 101.
Having to hike prices due to rising costs is understandable, and maybe they found it highly unlikely that low-cost clients would convert, but if you don't propose you can't retain. And more importantly the company has taken a reputation hit, very unprofessional.
The utility of credentialism for professional positions lies in the cost of required skill assessment Vs. the size of the talent pool. You can afford a fast filter like a college degree if you have a surplus of candidates. Especially if it generally is good indicator of cognitive ability, which OP admits it probably is.
A small pool of talent for niche requirements, will require significant adjustment and resources in assessing each individual that applies.
Some of the most hardworking people where I worked had a completed degree. Filtering people because they do not have X degree of Y years of experience is not right morally, but this is understandably done out of practical concerns. Nobody likes it but talent acquisition is expensive enough as it is.
Yeah, synonyms exist. If you get annoyed by technical terms having different meanings in different fields, you are going to have a frustrating ride through life.
There really should be a keyboard-shortcut standard. Some sort of configurable mapping of common actions/intents to key combinations that is set by the OS and shared between all apps.
I agree that I would love a discussion by OP acknowledging previous efforts and comparing approaches.
Having multiple minds working at the same problem is great, but not engaging with prior art diminishes the value of their own work IMO.
CarpalX [1] is also a keyboard modeling approach that includes typing effort and corpus loading. I would love to hear about the difference in typing effort modeling and approach for CarpalX.
CarpalX is a linear parametrized model with configurable parameters and weights for different types of strokes, so you can weigh the effort. You on the otherhand have made measurements of your hand movement. Very impressive and valuable! In the explainer [2] you mention the datasets briefly skipping over specifics:
- Are the hand-movement metrics a measurement of the center of the hand from an video-based object detector?
- Do you have individual finger detection too (doubt it is very precise due to occlusion)? My guess is relative hand movement is a good approximation of overall finger movement though.
- Any typing effort model based on this data will not be applicable to ortholinear or alternative layout boards such as a split hand (Ergodox, Corne, Moonlander or Plank).
- You briefly mention removing any manually determined effort-based objectives (like those in CarpalX) from the genetic algorithm optimization. You say they are highly similar to the purely trigram data-driven approach, but I am still very curious to see the results with these objectives included.
I am not trying to put you down though: It is still very commendable to research a better layout for the most common staggered 60% lay-out. I have been keylogging myself for three years now and will definitely try out your approach though, looks very promising!
Great to see as much thought is put in infrastructure as in the improved cryptographic logic. They removed the /dev/urandom and the virtual machine cloned entropy that cause unsafe randomness.
Alongside BLAKE2 algorithmic improvements, we also get safer infrastructure. Very cool!
So this is allowing remote git repositories to be treated as a regular filesystem. I really cannot think for good use-cases for this.
Maybe to use existing file-based indexing and search tools? I must be missing prime use-cases here.
It would have taken you as much effort to read the "shader" wiki entry [1] as it did to pendantically put down a foss library website for privacy issues. You would also have learned a fundamental concept I 3d graphics.
Great to reduce the complexity of shader programming and making it more accessible to front-end devs. Shader elements can lend a unique artistry to web design if not overdone.
I haven't looked very deeply in the codebase but it seems to be a JS abstraction/generator library for GLSL. So now the high-level pipeline as I understand it is ShaderPark > GLSL > WebGL.
site design / logo © 2022 Box Piper