HN user
ex-Github, now making software I want. Interested in cars and personal finance. Writes at https://jch.github.io
It's frustrating not being able to send email for my hobby projects even if I follow all the rules and have the correct headers. I enjoyed reading jeremyevan's post on self hosting email, but it's only for receive and not send https://code.jeremyevans.net/2021-07-29-running-my-own-email...
It changed the cost/benefits for papercuts. Instead of thinking, that's annoying, but not annoying enough to fix, I can spend a few minutes running things specific to my configuration. Fixing bluetooth, fingerprint readers were recent ones.
My last use case for it was selling a car and giving away some free stuff. Sadly, those have been replaced by fb marketplace.
Love rockauto, in the same vein, but not used daily is https://www.mcmaster.com/
This won’t happen, but I would love an Apple implementation of the Samsung Dex - phone soc that can dock into a laptop or desktop enclosure.
Lot of nostalgia today. This brings me back to compiling Camino, a gecko based browser, on my iBook G4.
I used to have a floppy and a mini-cd boot version of these. The mini-cd looks like a credit card and fit into a standard size cd drive. Reading the history of the project is a bit of a bummer, but still love the project ethos.
Correct and performant way to calculate historical value of a portfolio. I want a pure function, but taking a date as input is insufficient because users can edit holdings, and securities can split.
Weighing the tradeoffs of doing this calculation server or client side. That'll be an architecture shift away from my current set of background jobs fetching state and towards something more functional and on-demand.
I was just chatting with a friend how it hadn't changed in years, but there's a new beta when I opened it up.
- dark mode
- 3-column responsive layout, interesting it hides the watchlist
- like the view transitions for expanding content
- improved charting
- removed portfolio and holdings. Likely unpopular, but I thought it was a good way to track investments across banks without a 3rd party aggregator
I like reading to my kids and try to read to them in English and Mandarin. My Chinese is conversational, but I have a hard time finding books for them because I’m not good at writing. Something like this with language learning tools would be awesome.
I also like making up stories when we go on hikes. Long, rambling stories about unicorns befriending spiders and flying to faraway lands.
Like the layout tiles you have for the photo thumbnails. Will dig through and learn some css. Have struggled with different size content to create a compact masonry layout.
Living in hongkong for a few months, and absolutely love exploring the different neighborhoods. I’d love something like this or walkscore but for local guides to contribute.
Radio is so much fun to learn. It’s liberating to learn for curiosity and joy rather than commercialization. The community is welcoming, and while not directly translatable for most paid work, it does teach general problem solving skills.
Financial independence tracker. https://jch.app
Redesigning investment holdings for wider screens and leaning on hotwired turbo frames. Thankful for once-campfire as a reference for how to structure the backend. The lazy loading attribute works great with css media queries to display more on larger viewports.
Enjoying learning modern css in general. App uses tailwind, but did experiment with just css on the homepage. Letting the design emerge organically from using it daily, prototype with tailwind, then slim it back down with plain css.
From a PWA, I'd prefer a push notification, but you'd get a different answer from each person you ask.
Looks good! Appreciate that it’s a pwa instead of an app because it’d be something I use rarely. My issue isnt with the app, it’s that I’d have to remember to upload a receipt and also to find it in this app if I needed it again.
3 years. Left work because I felt comfortable with finances and family life, but wanted to try something different. It’s been fun working on personal projects and sharpening old tools. Still figuring out what I want to do long term. Some ideas include becoming a CFP because I like helping people with their finances, working for a tech company in that domain, or expanding the personal project (jch.app) and building more community.
In a good headspace now, right after the first year was feeling lost on where to go next.
Ya, no need comments or updates to it either. But thought it was an interesting piece of web development that's come up in every web framework and server, but has not been pushed into the browser. Would be interesting as a js polyfill even.
Thanks for touching on accessibility. I believe that good accessibility leads to a better design. The nice thing about starting with HTML is there are good defaults.
I wrote another post about building a search form about progressively enhancing a search form (https://jch.github.io/posts/2025-01-30-building-modern-searc...). Starting with semantic <search> and <input> elements gives sane browser and screenreader behavior.
Perhaps my title came on too strong, but I'm not advocating against javascript. It's more about understanding capabilities HTML and CSS can handle, and what is better suited for JS.
The guide is concise and well written. I like how it builds in the same sequence I blogged about: HTML, CSS, JS, and introduces each in a small way with links for further study. I like how it starts with a static site, but I think the end of the guide could include an example for server side processing. It was clever to use 3rd party example like search with fetch, but an example with a small JS backend to process a form and persist to a database / file would tie it together. It sounds like server-side is out of scope for the framework, so perhaps a few examples of small backends and how to plug in?
When I was reading the header/footer section, it reminded me of this issue to push for first-class includes in web platform https://github.com/whatwg/html/issues/2791
Your Rails backend, on the other hand, will experience far more changes and API instability in the long run
I see this as different layers of stability. On the bottom is a solid foundation of web standards that's widely adopted and resilient to breaking changes. Layered on that is the web framework and language. Before ruby, I really enjoyed perl, so I've experienced the collapse of a language and community. Matz had some good insights drawn from the history of other languages https://www.youtube.com/watch?v=4MM5b2F9zrM
Javascript is fascinating because of its evolution. Its initial popularity and wide install base guaranteed a lot of resources for it to grow. I'm excited to see whether wasm can bring more languages to the web.
Left out of the post, but the stack is rails, turbo, and stimulus. Hosted on a 2016 MacBook Pro with a dead battery in my closet
Your explanation in this post makes a lot more to me than your site.
Natural language as source code: Your intention becomes the program
Reminds me of Cucumber testing framework
- I liked the 'surprise' button because I wasn't sure what to type in the search box. I'm so used to typing a location by default. - I like your examples in this HN post and also the categories below the fold. - I like that I can favorite things without registering an account. I've been experimenting with guest accounts and blank states on my site. - When I search for multiple words from the search tab, acts funny. "big windows" became "big". I tried a refresh and the autocomplete said "big windows", but submitting with enter filtered it to "Bath, UK". Maybe there's some timing issue here with fast submits? - I prefer the screenshots / pitch you have on the app store over the website. I like the idea of browsing and filtering, rather than starting with a specific search. Again, that first point that this is something different than location search and helping me understand what I'm supposed to do.
Beautiful site, well done! I love the idea. I'm based in the US.
Appreciate the post and the discussion around it. As a car enthusiast, I enjoyed reading the specific implementations out in the wild today.
What stood out in the article was the attention paid to designing intuitive haptics. What our bodies expect in feedback when we turn a weighted dial. Doing it in software allows controlling multiple things, but I would still prefer a dumb dial per function, placed in close proximity to where the function acts. Seat controls near seats, climate controls near vents. Personalization stuffed into the screen and menus and profiles.
I did not think I would like auto climate controls until I lived with one. Even then, I would fiddle with it because I didn’t like the sudden high initial fan speed until I got used to it.
Added functions:
Simple Clock Daylight percentage Sunrise/Sunset Timer Face (Advanced Countdown) Alarms (16 alarms) Stock Stopwatch Moon Phase Pulsometer Alarm Thermometer Toss Up (rolling dice/flip a coin) Time Left (Days passed since my birth/Days left to 90years old) Secondary faces
World Clock (Dual Time) Tally Counter Save/load Databank (Emergency phone numbers +important passwords) Preferences Set time
I enjoy reading the daily news fetched from calibre and synced to my kindle e-reader. Love the e-ink display, and the ux for reading and navigating sections is great.
This reminds me of Berkeley's CS61A when it was taught with Scheme. One of the projects was writing a schema interpreter for scheme. It felt silly, but was a great small project to show case recursion, trees, and blurring the distinct between data and code.
I still remember Professor Brian Harvey rolling out his terminal on a cart in Berkeley's CS61A and typing out commands in a scheme repl. Learning what a y-combinator was with Structure and Interpretation of Computer Programs, and building my own scheme compiler with scheme.