Feel free to tell what you plan on doing this weekend and even ask for help or feedback.
Please keep in mind it’s more than OK to do nothing at all too!
Finishing my thesis to finally graduate! Need to submit it this Monday.
Alright! Good luck, it’s great when it’s over!
I am building a search engine with my friend who just graduated from a conversion masters in computer science!
The goals of this are to:
Any related resources are appreciated. We’ve been reading Write an Internet search engine with 200 lines of Ruby code and also the PageRank white paper.
I collected quite a lot of links about search engines. I think it’s a cool learning project.
Ah, a fantastic resource, thank you!
I see you’ve listed some search engines written in Rust. One of my favourite Rust projects at the moment is a WebAssembly search solution for static sites called Stork: https://github.com/jameslittle230/stork
Does the search engine has any distinct goals or new features? I just thought of Sonic, which is a light weight copy of Elastic Search and very great :D
Nothing novel. Just the standard Crawling, Indexing, and Querying. We were thinking about making it ‘a search engine for technical blogs’. Thanks for the heads up about Sonic, I’m checking it out on GitHub now.
a search engine for technical blogs
a search engine for technical blogs
I came across https://hyperlog.app/ which uses redis on HN (https://news.ycombinator.com/item?id=24383857).
I’ll be trying mountain biking for the first time and also baking challah.
Gonna learn some Forth, because it seems very cool :)
A great way to learn forth is to either implement a forth, or read through the implementation of a forth. Jones’ Forth is pretty good for this if you’re familiar with 32-bit x86 assembly.
Adding better tag support to claw. claw is the Lobsters reader written in SwiftUI. Currently single tag story viewing is supported but I want to support multi-tag story viewing.
More Linux from scratch and work on some draft blog posts.
Just remembered it’s the Plan 9 bootcamp at SDF too :)
Getting back into my non-technical hobbies. This weekend I intend to continue teaching myself Latin, and finally sit down and plan out a novel that’s been brewing in the back on my mind for a while.
What resources are you using to learn Latin?
Right now, a pirated questionably acquired epub of Wheelock’s Latin. I’ve heard good things about Lingua Latina per se Illustrata though. I went with Wheelock because I prefer a textbook that is closer to a grammar reference than a first course in a foreign language.
It’s a dead language. I always preferred the teachers who didn’t try to pretend the purpose behind learning Latin was to have conversations.
Starting to playing Gloomhaven with three friends. No work!
I’ve heard SO MANY good things about this game! The board gamers I know are just GUSHING over it.
Last weekend’s work on display server-side text rendering for some clients payed off but was quite taxing. It’s now about 2/12ths on way to completion but the benefits are starting to show even at such an early stage with minimal caching between clients – memory consumption is way down and full-dirty terminal emulator refreshes are at microseconds with low jitter.
Might continue down that road this weekend, though I doubt there is enough scotch left in the cabinets for that. Other option is that I picked up one of those 4k60Hz HDMI to USB3 converters that I hoped to use with my microscope. Kernel panic on one of my boxes and on the other the UVC implementation was being retarded about format picking and colour spaces, so that’s a slightly more “fun” debug target.
Playing games with family and friends, enjoying a regular VRChat meetup, taking a break from Hades (fantastic game), farmer’s market, and reading.
Glad to discuss each of those in more detail.
Write about how chrono and regexes work in Rust. I may get a little bit into testing for this too.
I had a Python script that is getting data from Airtable, crunch them and send a report via email through Firefox Send.
Unfortunately Firefox Send is no longer working, so I had to migrate. I would like to migrate over Lambda + S3, so I can take advantage of serverless.
Changing guitar strings, and packing for a camping trip. As trips abroad or to see friends is not a possibility, we’ve rented a camper van and hoping to explore a bit locally. Oh, and maybe visit our new allotment to clear it and prepare for planting.
Just finished up a basement remodel so I’m moving my home office downstairs to my library. Should make for a classier looking video conference background (current office is a pretty bleh room).
I don’t know about you folks, but I find a pretty small change of scenery can really help me out of a the mental ruts I tend to fall into.
I’m… actually hacking stuff. My father wanted to come over for his 70th birthday and mine, but due to rising infection rates decided to not travel cross-country.
So, I’ll try setting up a Mozilla Hubs instance and have a birthday party on it on Sunday.
Started playing Myth II again, having installed it from the CD. Also, the standalone game Green Berets. I recall as a youth making fun maps with the units from Green Berets, but for the life of me I can’t recall how I imported them into Fear. There’s also Loathing. It’s all been a fun blast from the past. The save icons being floppy disks made me smile. So, I guess I’ll try to make a map with those units while rewatching Miracle Mile.
Redesign a FP app to be modelled based on FRP.
Yesterday I night I familiarized myself with some terminology from classical FRP, notably events and behaviours - and the relationship between them. Over the new few days, I’m gonna continue rewriting the app while trying to get a better understanding of (so as to be able to teach) the FRP model.
I have been using FRP for writing web apps for a while now (via obelisk), however using it for non-web apps is not yet a natural intuitive process for me.
Did you find specific papers better than others to describe the model? I remember Paul Hudak did write a few good ones, but would be interested in your findings.
I’ve got a friend who is learning to code and plan on pairing with them for a small project. Hopefully pairing will help them transition between the “learn the syntax” and “tackle a project” stages, which I remember being difficult.
Other than that, working on my project that matches logs with models you’ve defined to verify everything is operating as expected. The algorithms involved are incredibly fun. Almost at the point where I’m proud enough of it to open source it. I think I’ve settled on the name TrailGuide, since it’ll be able to tell you when and how something in your system got lost or went off-track.
Working on a game for a game jam
I’m planning on working on my wife’s e-commerce website for a while, and/or Battle for Seclusa. Also a small spot of work to finish up a doc for work.
Play with the kids hopefully a bike ride. Deadlifts. Put some finish on some window trim.
A few small(ish) things:
Not much coding, I guess. If there’s any time left, I’ll try actix-web.
I finally found a raytracing engine to plug into my ionospheric modeling site, courtesy of Dave Robbins K1TTT (warning, little bit of an essay follows).
A “ray tracer”, somewhat similar to the graphics concept of the same name, is a simulator that predicts point-to-point radio propagation by “casting a ray” from the transmitter until it hits the ionosphere, takes into account (a model of) the ionospheric conditions at that point, reflects the ray back down to earth, then lets it reflect off of the earth/sea to the ionosphere for another hop if necessary. With a little iterative adjustment, we can find a path that connects two points, and then say whether a signal will propagate at a given frequency or punch through into space, and what the loss will be at a given frequency, which is useful for planning.
Most ray-tracers integrate more-or-less simplistic models of the ionosphere, with just a few parameters. For over a year, I’ve had the idea to take an existing ray-tracer, tear out its ionospheric model, and replace it with calls to my own model, which integrates realtime data. With a little luck, it should make better predictions.
The problem has been, that every ray-tracer I’ve found is either written in really bad spaghetti Fortran-70, or really bad spaghetti 1980s BASIC. I like to think I’m a decent programmer, I know several languages and jump between them, and I don’t turn my nose up at “organically evolved” code that gets the job done… but I’ve been unable to even figure out where to start with codebases like VOACAP or MINIMUF, So, I let that idea sit.
But during a conversation with Dave he mentioned that he wrote a prediction program for contesters in the 90s that included its own ray-tracing engine, and it was done in Borland C. I took a look at it, and… it’s actually decent code. It’s not super advanced, it doesn’t calculate circuit loss or reliability, it just spits out a minimum and maximum usable frequency. But the code is organized, the ionospheric model is cleanly separated from the path calculations, and the latter was understandable enough that I was able to translate it into some working Python (using numpy) in around 6 hours, with only one major bug to be squashed the next day, and now I’m getting results that seem to be basically real.
Next step is to actually plug it into the rest of my system, so that people can request a run from the website, and see the results on a proper map — right now I have to kick it off with podman run on the server, which spits out a big text file that I can visualize with gnuplot… but without a nice world-map background or day/night overlay, which makes it hard to interpret. Making it spit out an HDF5 dataset instead that my existing map renderer can consume is not too hard. Creating decent web UI and trying to make everything relatively abuse-proof… that’s a bit more annoying. But with luck I can get something up this weekend.
I’ll be taking a bike ride on the Silver Comet. Maybe once, but hopefully twice.