News for hackers. Stuff that matters.
Re:Is Web3 anything? by tracyhenry on Friday December 10, 2021 @01:27AM (#29507001)
> Once submitted, that “contract” will be carried out exactly as specified. It is observable to everyone, debuggable, and non-falsifiable.

But a centralized database can also achieve this, right? The premise of the general public trusting a blockchain more than an established, centralized 3rd party (e.g. government) seems so feeble.

Re:Is Web3 anything? by ulrikrasmussen on Friday December 10, 2021 @01:39AM (#29507058)
Exactly. Case in point: GitHub. We trust it to provide consensus on the latest hashes that branches point to, and if you use signed commits, then you don't even have to trust it to ensure integrity. The identity used to sign is completely controlled by you, and you can easily move to a self-hosted solution if GitHub loses your trust. Distributed consensus mechanisms are not worth it if you just design your system to (a) be auditable and independently verifiable (b) support identities controlled by the users.
Re:Is Web3 anything? by Longwelwind on Friday December 10, 2021 @02:08AM (#29507249)
If I understood Web3 correctly, it would be exactly like traditional web, except that:

  * The source code of all websites would be open-source, and accessible to anyone. It could still be changed, but it could not be done secretly.
  * All the actions of the users of the websites would be public.
If we take social forums, for example Hacker News, it would mean that:

  * The algorithm behind the front page of social forums would be known. Administrators would have no way of secretly skewing it.  
  * When a moderator would delete a message, they would merely mark them as deleted in the blockchain, and they would merely be hidden in the UI that you use to see the blockchain. The message would still be accessible on the blockchain.
The biggest downside, though, is that there would be a fee to any action done on the website (posting, moderation actions, changing your profile, ...).
Re:Is Web3 anything? by nicbou on Friday December 10, 2021 @04:24AM (#29508026)
The other biggest downside is that anyone else can just copy your website and profit off it.

I already have a pretty hard time with SEO spammers rephrasing the content I live from and passing it as their own. I can't imagine people lifting the entire website and making a permanent copy of it, profiting from 4 years of my work.

That's for a 100% free website with no "secret sauce" besides hard work. I can imagine any other website might want to hold their cards even closer.

Then there's also the site owners being held responsible for hosting illegal content, and being powerless to delete it.

Re:Is Web3 anything? by wccrawford on Friday December 10, 2021 @06:13AM (#29508594)
Even if they aren't powerless to delete it, they'd still have to pay to delete it. I wouldn't be surprised if someone finds a way to blackmail site owners with that, forcing them to pay out the nose to remove content they didn't even create.
Re:Is Web3 anything? by _gf4m on Friday December 10, 2021 @03:54AM (#29507854)
I think a platform like this is as impractical as you are thinking. However, there are potential other things you can do by mixing centralized services with decentralized ones. You could create a forum like hackernews that utilizes traditional infra but restricts access to users that hold NFTs or something similar. Think of something like a Patreon type deal where access to the service is tokenized. I've been kinda thinking of stuff I could do with a NFT project I've been working on and it's a thought I've been bouncing around.

There's no way I'd want to create something that is similar to what you're thinking, because it's pointless as you are thinking. Nobody would pay for this. Also messages such as these have no reason to be on an immutable blockchain. But you can certainly create specialized services for people that support your projects using ideas like this quite seemlessly.

Re:Is Web3 anything? by meheleventyone on Friday December 10, 2021 @02:38AM (#29507412)
I’d also put the moderators and administrators having no way of skewing things as a negative. They do lots of good things for the community like removing flags on topics, merging topics and reupping topics that are interesting for a second chance. It’s a good demonstration that pure algorithmic feeds still need gardening.

And the fee thing is why none of this will work in practice so it’s a really big negative. Who on Earth wants to be nickel and dimed across the Internet for really over expensive compute and storage?

Re:Is Web3 anything? by Longwelwind on Friday December 10, 2021 @02:55AM (#29507512)
All those moderator actions you described would still be possible in web3. Moderators would still be able to up a topic, merge a topic, delete topics, ...

But:

  * The fact that the moderators have access to these mechanisms would be known by all users
  * The use of these mechanisms by moderators would be public. If a topic is upped by a moderator, it would be logged in the blockchain. No moderator would be able to do it secretly.
Regarding the fee, it's true that it's a big downside. To get a feasible use case, the devs would need to find a blockchain that reduces them.
Re:Is Web3 anything? by meheleventyone on Friday December 10, 2021 @03:22AM (#29507675)
Or they could just publish a log of actions, this isn’t such priceless information that we couldn’t just rely on trust like we already do in the current situation that seems to work well.
Re:Is Web3 anything? by Longwelwind on Friday December 10, 2021 @03:40AM (#29507768)
I don't understand the point of your comment.

You're providing a solution that doesn't fit the initial requirement of a system being trust-less. So, indeed, your solution works, but as you described, it requires trust in the administrators of the website. The idea of web3 is that you wouldn't. "Trust but verify", if you want.

This could be used for Facebook or Reddit (given that they solve or attenuate the issue of fees), where people might have less trust in the administrators.

Re:Is Web3 anything? by meheleventyone on Friday December 10, 2021 @04:42AM (#29508156)
I'm just saying that HN has no need to be trustless. I don't see the value of that as a user.

For stuff at the scale of Facebook and Reddit you'd also probably end up needing caching as the chains would be too slow so you end up needing a layer of trust anyway.

Re:Is Web3 anything? by ElFitz on Friday December 10, 2021 @03:20AM (#29507660)
> The biggest downside, though, is that there would be a fee to any action done on the website.

I’d also add keeping an eternal, permanent, public record of everything as a big downside.

Re:Is Web3 anything? by linkdd on Friday December 10, 2021 @02:27AM (#29507355)
> The message would still be accessible on the blockchain.

What about the right to be forgotten and GDPR laws?

Re:Is Web3 anything? by berkes on Friday December 10, 2021 @02:48AM (#29507478)
That is a tough issue.

A blockchain is essentially an Event Sourced system. And Right To Be Forgotton is a common issue in all event sourced systems.

Several technical and one non technical solution are common:

Encrypt events with per actor keys. Delete the key when user wants to be forgotton. Or encrypt just certain attributes in the same way.

Keep actual data outside of events but only include a link to the data and checksum of the data.

Avoid storing PII in the first place. Does your domain model really require an IPaddress or First Name?

All have tradeoffs, none are silver bullets. But it is a solved problem in Event Sourcing, so it is a solved problem with blockchains (if only new programmers new this and reused patters instead of continually reinventing them)

Re:Is Web3 anything? by dmitriid on Friday December 10, 2021 @03:23AM (#29507678)
> so it is a solved problem with blockchain

You're saying this:

> Encrypt events with per actor keys. Delete the key when user wants to be forgotton.

Blockchain is immutable. Which key are you going to delete? Or, wait, you're going to centrally store these keys?

Re:Is Web3 anything? by berkes on Sunday December 12, 2021 @02:41PM (#29532883)
> Or, wait, you're going to centrally store these keys?

No necessarily. Three alternatives are common:

Data is encrypted with a key only the actor owns. Works for cases where that actor is the only one who should access it (I Reveal My Attributes type of schemes)

Data is encrypted with a sum (combination) of the key from an actor plus some other actors have: only you and the recipient can decrypt. Works for cases like messaging or p2p data exchange.

Data is encrypted with a OR of the key from an actor plus one by a trusted proxy: proxy and actor have ability to decrypt. This is closest the "centrally" storing. With the difference that "proxy" allows for a little more flexibity -in theory-.

Re:Is Web3 anything? by meheleventyone on Friday December 10, 2021 @03:24AM (#29507683)
The blockchain is decentralised though so you can't do things like per actor encryption that can be deleted AFAIK. It requires a separate, mutable source of data.

Similarly storing all the data off the blockchain just means you didn’t need a blockchain.

Re:Is Web3 anything? by jen729w on Friday December 10, 2021 @01:37AM (#29507048)
Isn’t it the visibility thing? Yeah a centralised database can do this. Good luck getting access to it. “Security” says you can’t.

I think that one of the allures of the blockchain is that it’s publicly visible. This may or may not be a fallacy of reasoning but hey, we’re human.

Re:Is Web3 anything? by newaccount74 on Friday December 10, 2021 @01:59AM (#29507197)
A lot of government databases are publicly accessible.

For example, in Austria we have the "Grundbuch", a register of land ownership. Everyone can inspect the Grundbuch database. Authenticity is guaranteed by requiring a public notary or a court to confirm any changes in the database, and there is a paper trail for every change to a record.

It's boring old tech, but it works, and doesn't have all the gotchas of blockchains.

Re:Is Web3 anything? by berkes on Friday December 10, 2021 @02:51AM (#29507491)
Many people complain about transaction fees and limited throughput of a blockchain.

I can only imagine this ancient system to be a magnitude worse in those regards.

Re:Is Web3 anything? by newaccount74 on Friday December 10, 2021 @04:35AM (#29508096)
The system is optimized for security, preventing fraud is more important than throughput in this case. Registering something in the Grundbuch is fast enough for property transactions.

Transaction fees are high by design. They are a form of tax on property transactions that discourages short term investments.

The system works very well and I just don't think it needs fixing.

Re:Is Web3 anything? by berkes on Friday December 10, 2021 @08:26AM (#29509539)
> The system works very well and I just don't think it needs fixing.

For the few use-cases that match it, yes.

The name Grundbuch itself puts a limit on its use, but we can imagine that there are many situations that call for the same amount of protection and trust, but which require a far higher throughput.

Back when Grundbuch was set up, ground and estate was probably the most valuable asset peopel can own. Today we have stocks, bonds, contracts, savings, companies, art, private-keys, hell; even cars, that are worth a multiple of a typical meadow in the Austrian hills.

Many of these classes cannot use the a system like Grundbuch because its fees are too high or its throughput too low. No-one wants to pay upwards of €600 just to sell some AAPL stock. No insurance company wants to drive to a notary for each contract they sign and then pay €600 upwards.

I'm not trying to say that blockchain is a solution for all these cases (It isn't). I'm saying that the system is limiting, problematic and unfit for many cases. That therefore, this system is unfit for the majority of cases that would need them 2021 (but for one specific case it works).

Re:Is Web3 anything? by nicbou on Friday December 10, 2021 @04:27AM (#29508047)
Depends on the system. We served webpages quickly 20 years ago. Berlin also has an open data portal, and it's as fast as any other website.
Re:Is Web3 anything? by hansonkd13 on Friday December 10, 2021 @02:24AM (#29507333)
Same thing exists in the US. Property ownership and property tax is open and publicly accessible
Re:Is Web3 anything? by ssss11 on Friday December 10, 2021 @01:40AM (#29507061)
Well not exactly. The blockchain is distributed and you effectively can’t modify it (let’s not get into 51% attacks etc)

A centralised database is controlled and modifiable by the owner - government, private company or whoever. They will act on their best interest not yours.

I don’t personally have a problem with government having my dat although they’re usually useless at implementing technology, but a decentralised blockchain would be way more trustworthy than private companies..

Re:Is Web3 anything? by aforwardslash on Friday December 10, 2021 @05:44AM (#29508446)
Blockchain is not necessarily "distributed", in the sense that multiple competing parties have a copy of it. As an example, hyperledger fabric is commonly used as a "private" blockchain.

Also, it is trivial to implement the same verification mechanisms in traditional databases, without the performance penalties of "blockchain", and transforming them in immutable, append-only storage systems that can't be tampered transparently. A good example of this is google trillian, the backend used by Certificate Transparency implementations.

Re:Is Web3 anything? by iamwil on Friday December 10, 2021 @10:28PM (#29518146)
A centralized database is not observable by default. It’s usually not directly connected to the internet accessible by anyone.

Databases as they’re usually designed today are updated in-place so they have no provenance of the data.

Blockchain data one written will exist in the history and cannot be changed. Unless specifically designed this way centralized db doesn’t work this way.

Re:Is Web3 anything? by tracyhenry on Saturday December 11, 2021 @12:21AM (#29518663)
> no provenance of the data

This is hardly true. It's easy to keep a history of query logs, which is enough to figure out 90% of data lineage. The rest 10% probably correspond to operations not even runnable on a blockchain.

I can mock a web3 application by opening a database to the entire internet. All edits logs will be visible, although edits will be access controlled. Now I'll claim this is observable. Some people would probably complain. But how in the world is the general public going to complain?

Re:Is Web3 anything? by drdeca on Saturday December 11, 2021 @04:07AM (#29519577)
I still think it could be potentially useful to, with a centralized database, use a similar kind of authentication scheme, in a way that allows for succinct demonstrations of any kinds of rule violations (compared to the rules that were set), as well as efficient to check proofs that state changes are in accordance with the rules which were set.

(I wonder if a network of such centralized DBs run by different parties might be enough to produce incentives among the different DB owners to behave in accordance to rules which they initially publish for themselves, etc. )

Re:Is Web3 anything? by errantmind on Friday December 10, 2021 @01:46AM (#29507112)
Feeble in what sense? Even absent blockchains, I only trust the people running the government in so far our incentives are aligned, and they aren't all that aligned at the moment.

Perhaps alternative structures of implementing governments should be explored.

Re:Is Web3 anything? by maccolgan on Friday December 10, 2021 @03:10AM (#29507601)
A centralized database can't achieve being public without putting trust in a third party (and no Web3 developer seem to eminently trust the government) to verify the double spend problem, otherwise blockchain is just a database. Non-falsiability can be verified through cryptography, yet we have no centralized but non-falsiable database system ready to use.
Re:Is Web3 anything? by aforwardslash on Friday December 10, 2021 @05:46AM (#29508459)
Well, you mention double-spend problem, that is a very narrow scope of applicability for a centralized database.

There is a huge public, cryptographically secure, append-only, verifiable centralized database (in fact, they are several), used by Certificate Transparency. Spoiler alert: Its a sparse merkle tree in MySQL.

Re:Is Web3 anything? by mattwilsonn888 on Friday December 10, 2021 @03:09AM (#29507593)
"Seems feeble?" You can go on about implementation details, UX flaws, problems with the state of crypto today - but what's been mathematically obvious since Bitcoin's inception is that the right ledger is both permanent and open enough to disregard issues of trust (you could say it has infinite trust, which is an exaggeration, but in a practical sense this is how its used without blemish), and that governments and central parties are easily corruptible and capable of cheating in ways so as to never get caught or noticed.

You either know very little about cryptography and game theory or very little about government to be asserting such doubt. Crypto-currencies have many problems, but finding ones where trust can be taken for granted is trivial.

Re:Is Web3 anything? by timdaub on Friday December 10, 2021 @04:05AM (#29507918)
How can a database achieve observability to anyone?
Re:Is Web3 anything? by tracyhenry on Friday December 10, 2021 @01:54PM (#29513914)
just make it openly accessible...