Even in very specific instances the smart contracts can only ever observe the Information they are given and have to assume that all that information is correct. What if the donation was done fraudulently or in error?
These systems have no way to undo these transactions (by design). They simply move all the "error handling" or "fraud prevention" to outside of the system.
And yes: if you can pretend that errorsor fraud don't happen, then one can design much simpler sysfems
But those assumptions don't make errors or fraud go away.
Edit: another aspect that the "traditional Systems" have at least some provision for is to prevent abusive or one sided contracts from being entered or at least enforced. For some the lack of those safeguards is a feature. For me it's terrible. Tons of contracts happen between unequal parties so the law has to protect the weaker one more.
I promise this isn't a generic anti-crypto rant, but rather a specific anti-crypto rant:
There are many projects in this space that try to replace what they perceive as flawed legal systems with perceived "perfect" (or at least better) digital, automated systems.
And I definitely understand that urge: there are many problems with various legal systems ranging from annoying (like being slow and very disparate around the world) to massive (biases, lack of access for those who need it most).
So aiming to improve that situation is understandable. And being pessimistic about the chances of fixing those systems with the "normal approaches" (i.e. politics) is equally understandable.
Where these projects usually break down though is that they generally lack an understanding of what makes legal systems so hard to get right: no one has found a reliable way to encode a non-trivial part of the law into something that a computer can decide reliably and without wrong decisions. (there are of course other difficulties, but this is the most lenient one for the current topic).
People with a technical background (which includes me) are often frustrated how laws and legal documents like licenses are at the same time both written in an arcane inaccessible language and also very much prone to interpretation. We assume, based on the languages we interact with, that a sufficiently complex language should allow a strict, formal interpretation of some truth value ("was this contract followed by both parties?").
But the reality is that contracts (just like most laws) are intentionally written with some subjective language to both account for real world deviations and avoid loopholes.
It's incredibly easy for a law to apply when it's not meant to (or the opposite: to present a law as not being meant to apply to a certain situation when the authors were very aware of the implications) or to not apply due to some technicality.
And for all the wrong in legal systems that exists we have not yet found abetter way to solve this than (hopefully neutral) arbitrators that interpret the text and underlying intentions.
And all the crypto schemes categorically decline that: their stated goal is to not have a human in the loop anywhere. That would be fine if they also solved the above problems in some other way, but none that I know of even attempt to do that. They simply pretend that perfect, decideable contracts are possible (even easy!) and never unfair.
Whether that error is based on ignorance or on something more sinister is up to the reader to decide.
Seconded! I own a domain for our emails (no public Web presence) and use a subdomain (that's not publicly hosted, the names only exist in my pihole) which allows me to use foo.l.mydomain.com for each service. Since the names don't resolve publicly you'll have to use dns verification for let's encrypt, but that's not too hard to do.
I know that I would keep forgetting to update the docs, so my documentation are the ansible playbooks and docker-compose.yaml files that I use to set it all up.
That leaves anything that has to be done in some Ui undocumented, so I try to keep that to a minimum, which isn't always easy (I'm looking at you authentik!).
And even if you absolutely want to self-host the serving of static pages: Making that secure and not prone to security issues is much easier than something that can actually execute PHP.
Note that the last hub inspected in the Article is a Anker product which also seems to be rebranded one from another third-party producer. It has some better components, but nothing fundamentally better about it.