Re:Web3 architecture and how it compares to traditional web ...
by acdha on Sunday October 10, 2021 @12:46PM
(#28819819)
> we basically stopped inventing shared protocols after that point in the 90s.
This is incorrect but educationally so: there are many new distributed protocols (RSS/Atom, XMPP, various things around blog interactions like FOAF and ping backs, GraphQL, OpenID, etc.) but the ones which successfully saw adoption were adopted by a few large companies who saw advantages from doing so. Being open or distributed not only wasn’t enough to be competitive with the centralized platforms, it often prevented that since you needed to coordinate improvements across a range of competitors while Google or Facebook could just ship an update.
Consumers didn’t value client decentralization, especially when that meant incompatibility, lower performance, and fewer people they know using a service.
Putting things into a blockchain not only doesn’t help with any of the real problems but makes all of them worse: performance and reliability are completely out of your control, everything costs money (good luck getting users when the competition not only has better features, perform, and social graph but also doesn’t charge them to sign up or send messages!), and since you centralized onto a network run by finance companies the only way to get traction on any of that is to have enough money to be taken seriously.
This is incorrect but educationally so: there are many new distributed protocols (RSS/Atom, XMPP, various things around blog interactions like FOAF and ping backs, GraphQL, OpenID, etc.) but the ones which successfully saw adoption were adopted by a few large companies who saw advantages from doing so. Being open or distributed not only wasn’t enough to be competitive with the centralized platforms, it often prevented that since you needed to coordinate improvements across a range of competitors while Google or Facebook could just ship an update.
Consumers didn’t value client decentralization, especially when that meant incompatibility, lower performance, and fewer people they know using a service.
Putting things into a blockchain not only doesn’t help with any of the real problems but makes all of them worse: performance and reliability are completely out of your control, everything costs money (good luck getting users when the competition not only has better features, perform, and social graph but also doesn’t charge them to sign up or send messages!), and since you centralized onto a network run by finance companies the only way to get traction on any of that is to have enough money to be taken seriously.