I'll start with welcoming a noticeable crowd of new readers to the newsletter. Thanks for joining us. Happy to have you.

I'm back to running the business again from parental leave and happy to put these idle hands to work. On the business side, the attention from Hacker News on my post The best parts of Visual Studio Code are proprietary has driven increased interest in my mentorship efforts which is heartening. If you are interested, the best way is to reach out directly, just respond to the email here and we're off to the races. I'm working on some more clearly scoped options to offer directly on the site as well.

As for development work, I would love to get in touch with your business if you have a need for reinforcements in Elixir, Python or even some Javascript (shocking, I know).

I'm also working on some more writing that is bound to show up on the site eventually. I also just did a thing which involved a screen recording, two people and a video camera. I have no idea how long that will take to edit but it'll show up in the RSS feed when it happens.

That's the business end.

Podcast recommendation: Automators

This podcast is more interesting in the wider concepts than the specifics. They cover the idea of automating workflows, for productivity or fun and making the machines do your bidding while mostly maintaining a very low technical threshold. They are however extremely bought in to the Apple ecosystem, so fair warning. I find the ideas and thoughts useful but prefer scripting things myself to the extent I automate.

With automation I find it fits better in technical workflows than almost any other workflow, specifically since you are already in the belly of terminals and technical tooling. So the step to get a script running as part of it is usually fairly simple.

I find the two hosts charming and friendly. Rosemary works as a developer and David is part of the wider Apple illuminati of podcasters.

You can find the podcast via Relay.fm or by searching your podcast app most likely.

Oh, the software potential!

As I might have tipped my hand to in previous posts such as Artisanal software - Beyond pragmatism and More than one thing at a time I have some thoughts about software quality, trade-offs and how we choose to build things today.

Take messaging applications for example. If you want to create a place to have conversations for a group of people that you don't know there are two very common answers: Slack and Discord

Slack is designed to be a business tool and has no focus on community moderation or group management beyond "of course people will behave, it is their workplace".

Discord is designed for communities and leans heavily on Gaming branding and can feel incredibly awkward in any professional context.

But that's not my primary frustration with these options. They are Electron apps on the desktop. They are disproportionately resource-hungry, they integrate poorly in the OS and they are basically unavoidable.

Discord has a lot of technical merit when it comes to voice calls. They basically made voice groups while gaming several times better than it was. That is what made them popular to start. Slack became popular. For some reason. I guess people like chat? Its fine, I don't hate it. And I guess the UX and design was their thing. I don't know that there are any outstanding technical differentiators in what Slack does. Aside from scale, they both deal with enormous scale at this point.

My point is, they are not best in class applications for anyone. They are same-everywhere and never-more-than-OK applications. The best they will ever aspire to be is capped by the fact of their developer-friendly and user-hostile cross-platform nature. And they do not aspire higher. They will not respect your resources, because who can be bothered to really complain?

I was looking around for options to this for small groups and communities and from an interview on Elixir Mix that is not released yet I was reminded that Telegram exists. I didn't know that anyone used it. But it turns out, it has a native MacOS client and Qt clients for all other desktops. Just this fact made me go "weeeeee!" internally because I knew that they would really have to screw up to not be much more efficient and much more pleasant to use than anything built in Electron.

The Telegram platform doesn't seem to have a business model beyond "we received a pile of money from some guy, its probably fine" which is certainly worth considering. But similarly Slack and Discord handing out free accounts is hardly future-proof. Also, Telegram seems eminently automatable.

I really wish we could get back to more native clients. I remember when Amarok on KDE was a "heavy" music player (aside: it had a feature to queue up a Stop, so good, I miss that). I remember when Spotify wasn't weird and slow, when it was a good local music player in its own right. I think the idea that native is better mostly lives on in the Apple ecosystem and mostly on the Mac. And it is also eroding there. I also find that Apple is overall testing my patience and I'm likely setting up a Linux desktop system here soon.

The biggest hindrance is that the incentives are very backwards. Time to market, developer productivity and cross-platform compatibility are beating out software quality, efficiency, performance and putting the user first. And I get it from a business standpoint. But I hope there will be a long-term reckoning where much like privacy the pendulum swings and the incentives shift. I do think these products leave all aspirations to greatness on the table. They can be good. But never great.

For now, smaller and independent developers along with dedicated FOSS creators own this niche I guess.

As always, if you have questions, thoughts or anything in particular to say, just reply or send an email to lars@underjord.io.

Thank you for reading, I appreciate it.

- Lars Wikman