Sven Slootweg @joepie91@pixie.town

Speaking from many years of experience with software development, personally and professionally: There's really nothing special about the stuff that large tech companies do, actually.

In the vast majority of cases where it looks like they have some magical technology that's miles ahead of what's publicly known/available, in reality they're using the same commodity tools and techniques that you are using, they've just papered over the sharp edges with marketing / manual labour / UI design / etc.

Sure, they have more budget, more marketing teams, more developers, more testers (well, hypothetically anyway), more control over the market and so on. That is all true.

But ultimately you could totally build the majority of the things they build yourself, as long as you set realistic goals. (Whether you *should* do that is a case-by-case question, of course...)

The marker has experienced an unscheduled relocation event

Testing my Wayland implementation by repeatedly turning a monitor on and off 🙃

The problem with Wayland is really not the protocol itself, nor that it is extension-based (that is actually a very good thing!), nor some weird conspiracy theory about it 'competing' with Xorg or being some kind of 'takeover' (it's the same people developing it! It's effectively just the next version.)

The real, actual problem with Wayland is that some desktop environments started defaulting to its use before it and its ecosystem were at feature parity with the systems and tools people were already using. Some projects jumped the shark. That's it.

We can be critical of something like Wayland without losing all nuance and inventing conspiracies and doom that don't actually exist.

So there's an increasingly common argument in favour of licenses that prohibit using the software for evil, or other difficult-to-define restrictions - the argument goes that it's *good* that it's unclear from a legal perspective, because that scares off the people you don't want using your software.

While that is true, and I agree with the *spirit* of the idea, I think that that's overlooking the collateral damage of this approach, which has two main forms:

1. Your software will be legally risky to *any* kind of high-profile organization, *including* the ones doing good work, and so it will be unavailable to them too

2. More insidiously, it makes it very difficult to build on top of, limiting the benefit it has to the *desirable* users. I'll explain this one more below.

Building on top of someone else's software is usually a big decision that's mostly irreversible, you become entirely dependent on the upstream; you need a pretty large amount of trust in the upstream to make that kind of decision, as the future of your project (and all the work you've put into it) will hinge on it.

This is a problem especially in the context of disabled and otherwise margnalized folks who are trying to tackle difficult problems; they'll often have a very limited amount of energy, and will want to make it count.

That means that they are both a) dependent on building on top of other people's work, to reduce the energy that's needed to build a thing, and b) *particularly* badly affected if something goes wrong with the upstream, and therefore need an even higher level of trust.

Not only that, but those same marginalized folks are also some of the most vulnerable to legal pressure, including from eg. copyright trolls.

All this creates a situation where such 'shaky' licenses become a hazard; anything licensed like that may not be safe to build upon, and even if it is, their *own* project may get disregarded by others because it inherits the shakiness of the upstream's license - and they may well be targeting a whole different demographic that *does* care about this, even if the upstream doesn't.

The end result is that shakily-licensed software is not safe to build on, and so you end up severely limiting how many 'levels' of "building on top of other people's work" are possible with it - and that may sound appealing from the perspective of a 'dependencies bad' ideology, but it hampers the ability for marginalized communities to construct alternative systems and infrastructure more broadly.

This is why I don't like those kinds of licenses. Doing this on a license level all but guarantees that it is a threat only to the least privileged people, while the likely intended targets (governments, corporations) can mostly just ignore such restrictions anyway and get away with it.

If you *must* use such licenses, then please at least make sure you have an alternative solution to the question of "how are people going to be able to collaborate around this and build non-oppressive systems".

But really, there are probably better ways to scare off governments and corporations than a legal system that's stacked in their favour.