Agreed, but my point is with a centralized network the lowest common denominator wins. There is no reason you can’t have QoL features on an open network, and thusly let everyone have the features that they care most about.

Can you imagine what a shithole the internet would have been if email wasn’t federated an open? There is absolutely no way that whatever centralized bullshit would have spawned instead would already be either long gone or enshittified to the point of being useless.

Good for you, you have a short list of requirements out of a chat service and discord perfectly fills your niche. But different people have different requirements for chat, and they don’t align. And network effects force people who have differing requirements to use the service with the most users which sucks.

For instance here are things that I require from any chat service that I use that discord completely falls flat at:

• Ability to run it on my linux machine without using an electron client (npm is a huge mess of supply chain attacks and I refuse to run any software that is likely to contain dependencies from it)
• Ability to run it on my AOSP phone which does not have any google play services installed
• Ability to write software to back up messages without fear of a company changing their API and breaking my backup system

I hate that Google is exerting even more control on the internet with their TLD, but I don’t really think this attack is made all that much worse with .zip TLD. I can already bury a .com in a long URL and end it in .zip just fine like so:

https://github.com∕foo∕bar∕baz@example.com/foo/bar/baz.zip

Or even use a subdomain to remove the @:

https://github.com∕foo∕bar∕baz.example.com/foo/bar/baz.zip

The truth is most people don’t look much at URLs outside of a domain to verify its authenticity, at which point the .zip TLD does not do much more harm than existing domains do.

For mitigation, Firefox already doesn’t display the username portion of the URL on hover of a link and URL-encodes it if copy-pasted into the url bar. It also displays the punycode representation when hovering or navigating to the second example.

Yes, you are correct in that a single individual’s action will make no difference, just like your single vote makes no difference either. However if everyone does their part it can make a massive difference.

While your individual contribution makes no difference, you still should try to do your part. Yes, change takes work and a bit of sacrifice. Just like how it takes time out of your day to do research on candidates and go to the polls.

If you don’t do the work, it doesn’t make you smart or clever, it just makes you an asshole taking advantage of others.

There’s no way of knowing, which is the whole problem with their model and why a lot of us self host things in the first place. Even if they super duper promise not to use the data, they could be lying. And if they are actually true to their word today, that could change tomorrow.

I haven’t used any flatpacks, mostly because they don’t seem to have a good solution for running terminal programs. (Also I don’t like that the application developer chooses the permissions to expose rather than the user.

However, I have been using bubblewrap which is what flatpack uses under the hood to sandbox. This allows me to run both gui and non-gui programs, and I have the control of exposing the minimum required permissions that I’m comfortable giving an untrusted piece of software.

I seem to be in the minority here but I personally prefer using $ and # to denote root. I like this because not everyone uses sudo and might not even have it installed.

That being said, if you already have other commands that are using sudo -u ... to run commands as a different user then it might be best to just be consistent and prefix everything with it, but if there is only a few of those maybe a # cp foo bar && chown www-data bar is an alternative.

it has a nice working sync of connection profiles (even of ssh keys…encrypted!)

Sorry, but what on earth does this have to do with a terminal emulator? Something like this makes way more sense as a separate tool. It’s like if I was making a decision of what video player to use because it can sync my browser bookmarks.

The way I remember the order is that the parentheses around the link would make grammatical sense outside of markdown (the goal of markdown is to still be fully readable even when looking at the raw source).

For example if I were posting on a forum that didn’t have markdown support which one of these would make more sense:

1. You can find that on this lemmy instance (https://lemmy.world).
2. You can find that on (this lemmy instance) https://lemmy.world.

Option 2 makes no sense grammatically. Then you just need to use the square brackets (which rarely show up in non-markdown text) to denote the link range.

Alternatively, if you still have a hard time remembering the order, you can use reference-style links which make it even more readable outside of markdown rendered contexts (note that there are no parentheses in this version, nothing to get confused):

[Here is a link][1] and [here is another link][2].

[1]: http://example.org
[2]: http://example.com