[Nasrudith]

That is very not new - even transaction based decentralized databases technically date back to the 1980s - even if they received less practical use. The theory of databases are pretty old and practically ancient in computing terms.

Before blockchain was a thing much less a hyped one cloud computing used it and before that some databases practiced 'tombstones' for databases. The easiest way to maintain absolute consistency has been known to be not really deleting things but marking them as 'deleted' and leaving them to be ignored by default until consistency can be achieved. That was the approach when cloud computing dawned to make things scaleable and avoid the bottlenecks and load-balancing server overhead.

Transaction based databases are logically equivalent to an append only database. If you have a list of transactions that add up to the same resulting database state - with the option of more versatility if you don't constrain things to be operational order insensitive.

[oelmekki]

I see, thanks to you and @weego for historical context.

It makes me realize that git is probably some kind of append-only database too (and used for decentralization too).

I guess I can still thank blockchain for having introduced the idea to me.

[jrumbut]

The similarity goes deeper than that, given that git uses Merkle trees as does Bitcoin. There's hashing and in the typical use case there is even some distributed consensus process involved although often in git human decision making takes the place of an algorithm.

Some good discussion here: https://stackoverflow.com/questions/46192377/why-is-git-not-...

I think the water gets muddied by people who know very little and yet have an opinion as well as by people who know so much they can't see the forest for the trees (if you'll allow the pun).