4 ms·

IPFS isn't http. Web browsers as software are meant to access and render html obtained via http. Content addressing on IPFS is entirely different to how dns res

by zanny 6y ago

IPFS isn't http. Web browsers as software are meant to access and render html obtained via http. Content addressing on IPFS is entirely different to how dns resolution works and as such you shouldn't just glue on an entirely different name resolution system to existing software. You want an ipfs browser that is UX optimized for actually browsing ipfs content. Back in the day browsers supported alternative protocols like ftp but in practice those are largely abandoned to single function as an http browser. And what in the world would adding PGP to Firefox do? Its not an email client. Its not an IM client. A website that wanted to can already use PGP via wasm. Firefox is not an SMTP client, its an HTTP one.

zoobab 6y ago
"Web browsers as software are meant to access and render html obtained via http" HTTP is a slave protocol, where you need a master. It's time to retire it, and design the web with an uncensorable protocol that does not loose memory. The slave design of a web server need to go.
- dredmorbius 6y ago
  Can you unpack this, because there's much that may or may not be in what you said.
  - 6510 6y ago
    The goal should be to render HTML documents not control where they come from. The server/client model or the master/slave model is only interesting if the clients are mediocre computers with limited bandwidth. Think of it like managing a bunch of children. It's a great formula until they are old and mature enough. If you continue to control them as if children they effectively become your slaves. I have tons of bandwidth and disk space. This is all that is required. If you unreasonably insist I should also be the one to serve every request it's game over for me. Not by technical limitation but by force. Oh, I can rent my own web space? I haven't money for this and it has been tried before. Those sites are all gone. Everything is gone. I have tens of thousands of blog postings talking about stuff that doesn't exist anymore.
dredmorbius 6y ago
Early browsers were decidedly multi-protocol, handling HTTP, FTP, NNTP, SMTP, and Gopher, possibly others.

by zanny 6y ago

zoobab 6y ago

"Web browsers as software are meant to access and render html obtained via http" HTTP is a slave protocol, where you need a master. It's time to retire it, and design the web with an uncensorable protocol that does not loose memory. The slave design of a web server need to go.

dredmorbius 6y ago
Can you unpack this, because there's much that may or may not be in what you said.
- 6510 6y ago
  The goal should be to render HTML documents not control where they come from. The server/client model or the master/slave model is only interesting if the clients are mediocre computers with limited bandwidth. Think of it like managing a bunch of children. It's a great formula until they are old and mature enough. If you continue to control them as if children they effectively become your slaves. I have tons of bandwidth and disk space. This is all that is required. If you unreasonably insist I should also be the one to serve every request it's game over for me. Not by technical limitation but by force. Oh, I can rent my own web space? I haven't money for this and it has been tried before. Those sites are all gone. Everything is gone. I have tens of thousands of blog postings talking about stuff that doesn't exist anymore.

dredmorbius 6y ago

Early browsers were decidedly multi-protocol, handling HTTP, FTP, NNTP, SMTP, and Gopher, possibly others.