This. It’s not as simple to get it working as it is on non-free OS’s, but with rclone I can get on Linux pretty much the same functionality I get from (eg.) Google Drive on Windows, including have most of the drive with on-demand access (meaning files are not stored locally, but downloaded / uploaded as needed) with a few specific folders synced for offline use. Since it supports a lot of storage services, I suppose it shouldn’t be that different to set it up the same way with Proton Drive.

Not really new. It’s basically LCD without backlight. So, higher resolution GBC / GBA alike screen.

Apparently it’s not that the software is broken, it’s that the software being installed breaks Windows Update. There are reports from people that uninstalling StartAllBack, updating the OS, then reinstalling it back (renaming the install executable first) works fine.

As much as being affected by this is frustrating to me (though this is all happening still on the dev channel, so for me it’ll be a problem for the future), I understand Microsoft’s rationale here. They can’t be expected to support every third-party tool that can break the OS, and it’s known that both ExplorerPatcher and StartAllBack relies on many hacks using undocumented APIs to work.

In the last few decades that I’ve been using Windows, I never felt compelled to use shell replacements or customizations - the default experience always worked fine for me with a few tweaks. So, if anything I’m more frustrated at Microsoft that I’m forced to use StartAllBack, because MS went and removed options from the shell that existed forever and always took for granted, and then some.

Ah yes, the Bobby Tables approach.

I think there’s more to the coin ship, that unfortunately isn’t covered in the article. I know for a fact that the hammer bro on the World 1 map turns into a coin ship if you finish 1-1 with 55 coins AND end the level with time remaining being an odd number.

Because this looks and feels pretty much identical to the arcade.

Probably because it shares a lot of the hardware of the original arcade, so the porting probably was straightforward.

Chromium-based browsers still trounces Firefox on the Jetstream benchmark. I mean, I realize the Speedometer benchmark is supposed to test real-world scenarios, while Jetstream is more synthetic, but whatever work mozilla did to improve performance I’d expect to scale in other benchmarks too, so I’d expect Firefox to at least be bit closer to Chromium, even if losing a little.

Fair enough! FWIW, I also think your stance on the matter is fairly level-headed and well thought out, even if I’m more or less on the other side of the fence.

While I personally do not think that all Chromium browsers (especially since there are projects like ungoogled-chromium) transmit your personal data, I can’t verify this myself because the Chromium codebase is far too much of an undertaking for myself to review.

Don’t you think that, with so many contributors and projects having eyes on it (arguably more so than on gecko), if there was foul play wouldn’t anyone have sounded the alarm?

The closest thing to an explanation I could find online just said “legal issues”, but didn’t go into details.

I don’t think that makes sense, or at least it doesn’t properly qualify the problem. BIOS is a set of baked-in software routines that mediate certain operations between software and hardware. In theory it could be reverse-enginereed and thus emulated just like the rest of the hardware is. In fact, many of the more simple systems (like 8 or 16-bit consoles) have their BIOS emulated. But for more advanced or poorer documented systems, there are, in my view, two problems with that:

• If your reversed engineered version of the BIOS has bugs (and during early stages of development, it would have a lot), the ways in which these bugs could present themselves makes the situation ambiguous, because it may be hard to know, from the symptoms, whether the bug is on the BIOS or on the hardware emulation. So developers just use the official BIOS because then if you see bugs, you know for sure the problem is on the hardware emulation. And also, reverse engineering the BIOS would require a lot of effort that developers would probably rate as low priority given they could use a perfectly functional BIOS and avoid a whole lot of other technical problems as per above. I mean, for many systems, hardware emulation is a problem already complex enough;
• Depending on the system, the BIOS code could be so simple that a reverse engineered version of it could conceivably be so close to the actual official code that it could, yes, trigger a copyright suit from the creator.

Huh, interesting. I guess now I know where BeOS / Haiku got its inspiration for how their windows look.

Maybe I’m bitter, and I know a lot of people wouldn’t agree with this, but honestly? I think the non-corporate part of the Fediverse should just assume malice from the get go and preemptively defederate from whatever Meta put out. That way nothing’s changed - Meta would essentially have a private / proprietary / isolated network, as far as users are concerned (much like Facebook already is), and even if the Fediverse will see less growth in the short term because of that, there will be no confusion on where everybody stands.

I don’t know about Jerboa, but this happens on the web too, and it’s usually on more fringe / less popular communities. There’s a workaround that works for me, I mostly heard about it from someone else here on Lemmy. It’s not great user experience, but it works.

• Go to the search function;
• Type in the name of the community, hit enter;
• Select “Commnities” on the search type dropdown (yes, you have to do it after submitting the search as per the previous steps, seems like another UI bug);
• If after doing the above the community is not found, then here enters the magic sauce:
  • Change the search terms so the community is fully qualified / keyed as !communityname@instance.server, all in lowercase. So, for example, if you’re looking for the “kreisvegs” community on feddit.de, type !kreisvegs@feddit.de on the search field;
  • Submit. It still won’t be found. Give it a few seconds, change back the search terms to only the community name (without the instace and “!” at the start). Submit. It should now show up.

You may have to repeat step 4 onwards a couple times, sometimes.

My uneducated guess as to what’s happening: the federated communities listing is probably cached on the instance, and by default it’ll only look for communities cached on your instance. My guess is that federated communities only gets into the instance cache when members of the current instance have searched / subscribed to that community. Typing the fully qualified community name on the search field (which is the tip I got from someone else) apparently forces the search function to actually contact the external instance to look for the community, instead of looking in the cache, but that can take some time, hence why you should wait a few seconds on the 6th step. That guess could also explain the problem also happening on Jerboa, since the problem would be server-side.

I agree. It’s just people being lazy and resistant to change. The irony is that it was exactly those “big techs” that bred this kind of behavior on them, now they turn around to criticize those big techs, but have no spine to do what actually takes to change that, so they keep on taking the abuse.

I think what he’s getting at here is people asking him about building an alternative platform to Reddit - much like Lemmy, Kbin, Tildes, etc. are - as opposed to an app serving as the frontend for an existing alternative platform. I’ve seen poeple buzzing about that on some of those threads about Apollo on Reddit. Those people are so in denial that I couldn’t conjure to effort to point how ridiculous that idea is. Some people did, and were quickly shut down with the usual complaints (fediverse is confusing / does’t work, etc.)