Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A plan to rescue the Web from the Internet

staltz.com

Post reply on HN
Print view View original

11–20 of 220 posts

Re: A plan to rescue the Web from the Internet

#11

Post by xj9 »

whimful wrote: post #2

scuttlebutt is my main social media these days... oh and in case you're looking for git that's not wedded to github via your comments and issues, scuttlebutt plays really well with git - you push comments and code into your gossip-cloud together

plays really well with git

yes, but there are still some limitations that i've run up against (which i am trying to fix) when it comes to repos with a lot of history.

you can work around this issue by using a lot of small repos, but once you reach a certain threshold it becomes impossible to clone a repo from ssb[0]. going further, you also cannot push repos beyond a certain size[1].

these are sample repos that these claims are true for, but i still have more measuring to do so i can say what the exact threshold is.

i lean very heavily on the tech that i use to find the places that need improvement. that said, ssb is an amazing technology and quite new. it works very well for the social networking use case. with some work, we can make it fast/reliable enough for general applications.

[0]: https://git.heropunch.io/%25SI144KecoPoa4uB8kGGUbC%2Fb6Mb4eP...

[1]: https://git.alpinelinux.org/cgit/aports/

Re: A plan to rescue the Web from the Internet

#12

Post by JohnStrange »

ashark wrote: post #8
tannhaeuser wrote: post #4

I'm all for p2p (really), but how about taking back control (sorry) of the Web by actually developing and enforcing declarative/markup technologies and standards instead of praising JavaScript because "it's not half bad" and adding procedural features to the Web (APIs, WASM)? With the current state of affairs wrt privacy and self-proclaimed standardization bodies, I'm not sure the Web is worth preserving.

HTML with great default styles and better built-in form and interactive elements (date picker, sortable tables) would be ideal. Screw CSS, it slows everything down, just have good defaults and don't let pages mess with them much. If we must have CSS at least jettison the complex parts with animation and such. Make it very, very simple and weak. No Javascript. Can't be trusted if it has the ability to initiate connect…

I'd like to see a return to semantic markup, accompanied by pragmatic information about the purpose of the data and its intended size. The data should also be strongly typed. Then leave it to the client how to display, how to edit, and how to select the data.

The receiver should decide how to best display e.g. a list of 20000 names or a choice between 4 options on his device.

Re: A plan to rescue the Web from the Internet

#13

Post by nickbauman »

tannhaeuser wrote: post #4

I'm all for p2p (really), but how about taking back control (sorry) of the Web by actually developing and enforcing declarative/markup technologies and standards instead of praising JavaScript because "it's not half bad" and adding procedural features to the Web (APIs, WASM)? With the current state of affairs wrt privacy and self-proclaimed standardization bodies, I'm not sure the Web is worth preserving.

The vision of these three posts (tannhaeuser, ashark and JohnStrange) are incompatible. This thread is a tiny (but complete) clinic in why the web is such a mess.

Re: A plan to rescue the Web from the Internet

#15

Post by niftich »

The IPv4 perspective is a red herring. NATting was indeed necessitated by IP address scarcity, but a domestic installation that does NAT comes with ancillary benefits, like giving you, the home user, a single place to control access to your network.

In IPv6 it's nice that you have an address space that's not only big enough to accommodate every device, but large enough to even burn through addresses and treat them as disposable, but once IPv6 becomes widespread there will need to be some rethinking as to how to manage firewall rules between your own devices, how to segregate your portion of the network from the spurious (and sometimes malicious) traffic of everywhere else.

Re: A plan to rescue the Web from the Internet

#16

Post by EGreg »

romaniv wrote: post #10

The reason the Web needs rescuing is because it's not a particularly well-designed system that has been patched over and over again for the last quarter century. And now it has been degraded to a delivery layer for JavaScript apps, a poor RPC protocol and an overly complex UI rendering toolkit. It should have had protocol-level technologies for preserving historic data, for search, for offline use, for authoring and…

Yeah, the Internet had coalesced around one standard, TCP-IP, after fighting with national standards and networks like Minitel.

Now with the Web, the approach has always been to define the front end standards for "user agents" and leave the back end standards to "someone else".

Here is what would make the Web wayyy better:

Support for content-addressable protocols via IPFS ans SAFE. You can already leverage SRI to get some of the "content addressable hashes" but you need a DHT to locate and host the stuff. That way the Web becomes resilient (no single points of failure and you can trust the content) while retaining all its front end goodness.

No more concepts of DOMAINS at all, but rather, each resource of information can be an evolving community in itself. Like people collaborating on a document.

I think this could be the standard for a new multiple-readers-multiple-writers decentralized web: https://github.com/Qbix/architecture/wiki/Internet-2.0

We are working on a mobile browser and extensions for desktop browsers to do these things.

Re: A plan to rescue the Web from the Internet

#17

Post by gregknicholson »

There's a lot of white on this map: [World Population Density](http://www.luminocity3d.org/WorldPopDen/) — the hard part is gonna be getting data between Europe and New Zealand without Big Wire.

I love how Staltz's solution to this goes hand-in-hand with re-personalising our interactions. In short:

1. The Next Billion haven't yet become used to the idea that useful tech services must be global, and provided for you by a single corporation.

2. They want to communicate with other people they know, physically nearby-ish.

3. Sneakernet is never slower than talking in person.

4. Isolated mesh networks and comms with days of latency are viable for these new net users.

5. Get enough people using a mesh and the networks will start to connect.

6. Where the internet is already pervasive, privacy and autonomy advocates are resisting corporate control, and choosing decentralised alternatives.

7. For now, we can exploit the existing internet to handle long-distance.

8. But eventually, enough people will have the bottom-up mindset, and the weak links in the mesh will become worrying.

9. So a solution will emerge to fill in the gaps in the mesh, using ships / beacons / balloons / satellites / modulated whalesong.

Re: A plan to rescue the Web from the Internet

#18

Post by wvenable »

ashark wrote: post #8
tannhaeuser wrote: post #4

I'm all for p2p (really), but how about taking back control (sorry) of the Web by actually developing and enforcing declarative/markup technologies and standards instead of praising JavaScript because "it's not half bad" and adding procedural features to the Web (APIs, WASM)? With the current state of affairs wrt privacy and self-proclaimed standardization bodies, I'm not sure the Web is worth preserving.

HTML with great default styles and better built-in form and interactive elements (date picker, sortable tables) would be ideal. Screw CSS, it slows everything down, just have good defaults and don't let pages mess with them much. If we must have CSS at least jettison the complex parts with animation and such. Make it very, very simple and weak. No Javascript. Can't be trusted if it has the ability to initiate connect…

The problem with that approach is that you have to define all your possible use-cases up front. If the web had stayed as a simple document format, it wouldn't even be worth talking about now.

A much more flexible and future-proof approach is a fully programmable environment without any declarative features at all. No HTML. No CSS. Just a clean API in a tight sandbox.

Re: A plan to rescue the Web from the Internet

#19

Post by dec0dedab0de »

JohnStrange wrote: post #12
ashark wrote: post #8

Earlier quoted context omitted.

HTML with great default styles and better built-in form and interactive elements (date picker, sortable tables) would be ideal. Screw CSS, it slows everything down, just have good defaults and don't let pages mess with them much. If we must have CSS at least jettison the complex parts with animation and such. Make it very, very simple and weak. No Javascript. Can't be trusted if it has the ability to initiate connect…

I'd like to see a return to semantic markup, accompanied by pragmatic information about the purpose of the data and its intended size. The data should also be strongly typed. Then leave it to the client how to display, how to edit, and how to select the data. The receiver should decide how to best display e.g. a list of 20000 names or a choice between 4 options on his device.

Me too, except I would say strongly duck typed. As in there should be sensible defaults when possible.

In any case, the designers won this argument a long time ago. They want to decide how you consume their page, it is not a document to them, it is a work of art. How dare you want to display it differently from their vision.

Post reply on HN