[pudo]

> Put differently: it turned out that permissionless publishing alone was insufficient. We also need permissionless data.

Has the author ever heard of OpenStreetMap, Wikipedia, or a million other projects like them? I'm also not quite sure "permissionless data" is how I'd describe a database that stores tiny amounts of data at a massive cost. Pay-to-play is permissionless iif you have the money.

Sometimes the cure is worse than the disease.

[throway453sde]

Wikipedia database is managed by the wikimedia staff. You have the permission to edit it for now but can not guarantee it will be make available always. This is what the author meant.

[grey-area]

Opensea is managed by opensea staff, Bitcoin and Etheream are in a very concrete way managed by the devs. You absolutely cannot guarantee anything about access in the future for sites based on web3 - if anything it is worse as there are two gatekeepers, the site owner and the cryptocurrency (so for example gas fees might go up making a tiny transaction very expensive, or a currency might fail).

Nobody wants a truly distributed org because it is impossible for anything to get done, just as nobody actually wants to use a distributed currency because the costs outweigh the benefits, and the web we have is distributed enough, though of course it has flaws and is in some ways too centralised now, but I don't see how web3 as currently sold helps to solve that.

The biggest problems currently are around identity (SSO), and microtransactions, but neither of those are solved well by existing cryptocurrency solutions or things like NFTs.

[whywhywhywhy]

> Opensea is managed by opensea staff

Opensea is just a viewer and tool to interact with contracts. If your NFT is removed from opensea it could still be interacted with through another platform or a platform you make yourself (see notlarvalabs).

E.g the developer of a large NFT platform called hicetnunc had a temper tantrum and deleted their entire platform. But because it’s decentralized another platform (objkt) now provides all the same functionality with the same data.

Because of this you can see why there are such passionate people either side, a lot to gain but also a lot to lose if you’ve normally profited from owning data.

[grey-area]

That’s the theory, in practice the largest market wins, nfts are often urls to a particular service and opensea does things like this:

https://blockzeit.com/opensea-nft-marketplace-stops-hacker-f...

This is very much controlled by the intermediary (as well as the chain it uses).

[waffle_maniac]

Microtransactions will be solved by Brave/BAT. They have the largest user base which is very important for negotiating with publishers and advertisers.

[fredley]

> can not guarantee it will be make available always

True! However I trust the wikimedia foundation, and the fact that because it is a valuable dataset there are many copies of it. If they really did restrict access, anyone could spin up their mirror.

Most of the decentralisation arguments seem to rely on some fetishistically absolute version of the world where you cannot trust any single person or entity, and crypto/web3 is the only reasonable solution if that is true. I, however, have more faith in other people, and am fine placing trust—to varying degrees—in others.

[paulgb]

> some fetishistically absolute version of the world where you cannot trust any single person or entity

And even then, there’s still trust necessary to the extent that web3 things interact with the real world. You can have provable on-chain voting in a DAO, but if the DAO owns any non-crypto assets, you’re trusting a human to execute the consensus decision.

[igorkraw]

...and society to run the internet you depend on

[627467]

> I, however, have more faith in other people, and am fine placing trust—to varying degrees—in others.

That's one perspective, another is: privilege.

[mattwilsonn888]

That's a great example of decentralization, but expecting anything which gains value from being decentralized ala Wikipedia to be handled solely or even marginally by volunteers is unrealistic.

[throway453sde]

> I trust the wikimedia foundation

The board members change so it can go wrong.

> ..where you cannot trust any single person..

You are missing the point. Obviously society runs based on mutual trust. One can not ignore the fact that a few people or group in power can create great damage. Crypto makes it possible to create "trust by design" systems so that a small group can not do great damage or censor things that undermine their power.

[igorkraw]

No it doesn't. It just creates the illusion of this and allows you to build systems with different tradeoffs.

[denton-scratch]

> not guarantee it will be make available always

Untrue. You can download the entire database now, and there are many clones.

[lottin]

Doesn't wikipedia keep a history of all the edits, and aren't these available to wikipedia users?

[DaiPlusPlus]

They're just rows in a MySQL database. Traditional RDBMS for OLTP are built around the assumption that all data is mutable. An unscrupulous Wikipedia DBA could unilaterally delete new history rows before anyone downloads a hard-copy and then no-one would be the wiser.

On the topic of OLTP RDBMS: things are changing though: ISO SQL (and at least MS SQL Server) have added support for "LEDGER" tables that extend the Temporal Table system (...that we know and love...er...hate) with immutability secured with cryptographic keys. Basically, you get to have your own centralized blockchain-like table.... so long as you trust your DBMS (and DBA) with the keys.

There's two types: append-only and updatable (which is still append-only under-the-hood, as destructive DML operations are still just appended to current history, just like with git).

https://docs.microsoft.com/en-us/azure/azure-sql/database/le...

https://docs.microsoft.com/en-us/azure/azure-sql/database/le...

It's pretty cool, I'll admit - I've been wanting support for true "read-only" tables for ages - the only thing I'm dreading is having to put-up with poor tooling support for a decade or two (e.g. Entity Framework Core still doesn't support temporal-tables, and SQL Server still only supports SYSTEM versioning and not the (far more useful) APPLICATION versioning system, argh).

That said, I don't think Wikipedia would really want crypto-signed page edit histories: it would obligate them to host truly objectionable content (otherwise the hash references would break), and there's far too many sociopathic griefers online for that to not happen...

[throwaway744678]

I believe they do; you can download archives at [0]. It mentions that the edit history can be downloaded by looking for dumps with page-meta-history in their name at [1]

The same is probably true for OpenStreetMap et al.

[0] https://dumps.wikimedia.org/ [1] https://dumps.wikimedia.org/enwiki/latest/