uzername

I don't mind WordPress. My company doesn't use it for its marketing sites though, prefers larger more monolithic enterprise options. But it's not a startup either. From what I heard WordPress was not favored because maintenance was difficult (plugins, vulnerabilities, upgrades) and also there was uncertainty about gpl. A while ago some startups I worked were using webflow and instapage for their marketing sites.

Certainly an interesting take with the footer as the nav. It could work on information centric sites. It might seem strange working in a web app or a hybrid environment, where portions of the site lean towards app while others don't.

Looking at the comments here, that's highly skewed towards desktop use, guessing at the HN audience here, skilled with computer usage and decent handlers of visually dense information, I think most common users aren't like that. Many are using smaller screens, many are on mobile the majority of their computing time. When they do have larger screens, their glaze over when there's too much to look at (I think because they're not used to it, and too busy to ever get used to it, but either way). All of that to say, the hamburgers aren't for us, they're for secondary flows for common users.

This is how Mazda handles the screen. The knob has a satisfying tick on rotation and rebound when pressed. The car's a little older now and the screen os is a bit slow, but usually keeps up with input. Radio and navigation have hard button shortcuts. It's just a good system.

I appreciate this pricing model.

I was on a saas vendor call just today (unrelated product category) and we were hit with a similar number to your example, in the $20k-$30k range, with a fixed component and then per use seat scales pricing. There was no discussion of activity based usage. I think if we could have a fixed base price, and activity usage pricing that's capped, even if the cap is higher, that would be awesome. Now for us, we're looking to buy in stop maintaining out own internal fork and all the downsides that come with it, just to make a meager sso shim.

I wish you all the luck possible here, and hopefully making the model work well for you and the team.

Tangentially related to the file converters are yaml and json formatters that are very popular online. I've seen my team open one and just paste away. Who knows what these tools are doing, and even if it was safe one day, the next day it could have been compromised. I'm sure the org has leaked hundreds, maybe thousands of kb by now. At least we don't have PII or any production data. I have an alias that extracts my clipboard, uses jq to format, and puts it back into the clipboard (macos), as an alternative workaround for these web formatters.

I've seen this for many years, I think it's even been tries here in Twin Cities MN schools. Bus driver shortage and route consolidation of the last few years probably didn't help, but perhaps the "take the city bus" routine for high school students has changed the behaviors a bit. Getting off the school provided dependency could help shift the timings to what works best and not necessarily need to account for sports students.

I was wrapped up in pre-nft blockchain hype 2017-2019. It didn't work out. I learned that not every salient technology is ready for adoption or will leave a legacy.

My top few:

- practice reading code that is older you're comfortable working with, that you did not write and have no prior knowledge base for - designing systems that are easy to replace; no matter what tech comes along, new tech will come after, and we're not going to live in the world where X lasts for Y time, it will be Y/c (some fraction) - writing well, that's tailored to specific audiences, aided by whatever tools are available (spell check, grammerly, chatgpt, etc) - as far as tech itself goes, I'm hopeful about edge computing, where performance and power is near today's level, for an extended time

I see a lot of doomsdayism here. I recommend avoiding that, it's a tough and awful away to live.

I believe it was Snowpack.

An anecdotal story: I bought some Windows 10 keys a few years ago from a site pcmag "recommended," where they were selling Windows 10 keys for maybe $55 each. I bought two at the time. I kept the keys saved in my email. Fast forward to last November during a minipc build, I planned on using one of those keys. I tried it. It was rejected as a non-genuine key. I thought being written up on pcmag was a recommendation, but it was probably not a vetted source beyond being cheap. I learned my lesson and I will buy through Microsoft or through Microcenter (us brick and mortar store).

We deploy everything in kubernetes in the live environments, with databases in rds, s3 for objects, sometimes dynamo for other data types or patterns.

Locally there's docker compose for the APIs (Java, python) and usually postgres. It's a good observation, interrupting that workflow would be quite a pain.

I work at a pretty large corporation that uses docker desktop. The engineering team is in the low hundreds spread across the world. We happily pay for what we use and it's all great. We don't really use many features of docker desktop though. We have our own image space (ecr) and I don't think anyone is using or relying on extensions. I'm sure some team will eventually socialize an alternative. The reason we went through with paying for docker desktop was probably the lost time and productivity of transitioning off without due diligence dedicated on getting another option to work just as well. I don't know if our org gets customized pricing, but if it's $24/mo for us, that's maybe a 1/3 of an hour of the least expensive run rate engineer we have (worldwide, again), per month. That's maybe 5 hours of run rate time per year per engineer. I can see why our org bought in at the time and continues.

On a personal note, I was fine with the change, since it allowed personal use still with docker desktop. When Docker Desktop for Linux came out, I gave it a try on a clean server. Unfortunately, even on a fresh Ubuntu install with fresh hardware, the reliability of Docker Desktop for Ubuntu was awful, crashing every few days into a stalled state. I had to make a cron job to watch it and maintain it's uptime.

Oh, fascinating. I've never encountered that before. Thanks for sharing that use case and extension.

Page Visibility API is used in tools like react query, where it will show the page with the existing cached data, but refetch in the background upon a page visibility event. It feels natural for some apps to have this update-on-focus behavior, and it's nicer than polling with set interval.

Intl API is great, but I feel it's somewhat hamstrung because Node doesn't have matching APIs.

I recently discovered ncdu while troubleshooting a logging issue in one of servers. Something broke and syslog began to inflate unchecked and consumed the entire disk's worth of space. ncdu helped explore the system and find out where the file was. Great tool.

I went with the pihole in docker approach with a Ubuntu 22.04 machine. Pretty smooth, I had a few quirks with docker desktop for linux that might not show up otherwise. Definitely my recommendation too.

A few months ago I did try a LinkedIn search but their profile hadn't updated with new activity.

A few years ago, as fresh IC/Manager hybrid with a couple juniors, I found myself with a junior who wasn't growing well relative to other hires (same team, different team, hires from colleges mostly, some bootcamps too). The two of us didn't talk about performance or contributing or engagement enough. I raised some alarms with my management peers and my manager, but I think it was too late. We transferred them to another team, hoping another team (less independent, more team based), environment, industrial vertical, new manager would foster some growth. A few months later the company let them go.

About a year later, they asked me for a recommendation and I said sure, I will do my best. A recruiter called me, I tried responding via voicemail but no follow up ever came around.

I really struggled to identify weak performance back then. I gave everyone the benefit of the doubt. I also kept the topic at bay for too long. My regrets include my slow response, but my regrets extend to the unknown, if I helped ruin this kids career.

You really do need to ask how they're doing, how they think they're doing, and if you have any concerns, air them early so there's time to change.

Congrats on the launch. I saw payload here on HN a few months ago. Good seeing it again.

On the "Enterprise" sso approach, are your first party offerings just plugins? I've worked with various node cms recently, and for our needs we often see the only way to extend is by dangerously forking the upstream rather than hooking in through the plugin architecture.

The second aspect I wondered about, is why Mongo? It doesn't seem as "open" compared to a postgres, although maybe works better for your hosting platform offering. I wonder if an adapter to pg/json might be possible.

Thanks, and good luck with launch!

We use alpine based images extensively where possible. Most popular language ecosystems and tools have an alpine option. Alpine as a base image is nice because it's nearly blank, but easy to customize unlike distroless. Recently one tool did drop Alpine, most based on it's reduced performance, and that was envoyproxy.

I saw the update video yesterday and he mentioned the video scheduled for today. I hoped it wouldn't be a hiatus or a we can't finish Stormlight.

It was handled well.

Yes, Nest uses decorators and the reflect-metadata package to get information about them dynamically. Annotations/decorators in other languages are fine, but in JavaScript/Typescript I feel the risk of using them as a dead end feature. I'd like the proposal as much progress as possible, but it's been quiet. In the meantime, a Typescript Nest like system without annotations could be better.

Is this a bet against React or a bet for ecosystem variety and choice? I feel it's the latter right now.

It's a shame that the Tailwind docs recommend to use craco. I think that's a major turn off. In that situation, I rolled my own npm-run-all customizations of the default scripts for start and build. It's extra noise but less than the documented approach.

I like this approach slightly better than what scarfjs was doing. I first ran into that with react query a few years ago. It was a chilling effect for me at least and I was glad it changed.

The idea of this being an enhanced link shortener is a good metaphor. I think if were ever encountering an image or a download that's not going through a dockerhub or GH link, I might give pause and think about it a bit.

I'm not defending enterprise usage of packages, especially somehow if it's core to the business, but I don't need anyone to come knocking either.

TypeORM's velocity and/or lack of a 1.0 release is troubling. I've been trying to push nest recently to gain some consistency vs wild-west express apps, but TypeORM has been the rough spot. I can see the appeal of using something like the pg driver directly, but then there's all that data mapping.

In my CS undergrad, I took a few ethics courses, one in the CS program and one in another discipline (Mass Communication). I recommend something like this.

I also suggest looking into CS coursework you anticipate disliking. I had no desire to learn about compilers and languages, but I took a rudimentary introduction course and it turned out I enjoyed the topics significantly.

Our team's have been using Miro for this kind of collaboration. I was surprised that it felt better than using a whiteboard in person, because the cost of changing the already drawn boxes and arrows went way down.

For more technical whiteboards, I've been using excalidraw. I like presenting the "drawn" look to help imply that this is an unfinished idea and we're sketching the concepts out.

Inspired by the Overcast variant of this kind of tool, I built a very basic browser based version using canvas and various web audio/video apis. It was fun to build over a couple weekends, but it ended up not being usable broadly due to speed (runs in realtime linear) and browser limitations of file export types (webm in chrome). If ffmpeg could reliably run in wasm, there could be alternative approaches. I concluded after I built it, I should make a headless non-browser version and it would be more usable, but haven't gotten around to it.

I'm still hoping, against all the odds, of a Mac Mini Pro or a Mac Pro Mini. I don't need a Xeon, but I don't want a thermally constrained chip either.

I have participated in multiple types.

1:1 code review in person, sort of like pairing at that point. 1:1 remote via PR, 1:N in person in front of a screen or projector, 1:N remote where you need multiple approvals and different sections may interest different people, and N:M where a small team worked on a feature and request another teams review.

I think of review broadly as the remote/pr kind. But all of these are valuable in different situations.