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

Re-decentralizing the Web, for good this time

ruben.verborgh.org

Post reply on HN
Print view View original

191–200 of 295 posts

Re: Re-decentralizing the Web, for good this time

#191

Post by int0x80 »

hathawsh wrote: post #35
kodablah wrote: post #8

I think the part ignored by so many is the need to decentralize the computers into the home. I'm not talking meshes or shared resources. For the majority of use cases, we don't need distributed storage, compute, etc. Just start making these self-hosted "servers", "data pods", etc as easy to install as desktop software and make it clear that they are inaccessible when the computer is off. People that aren't already wi…

There should be a marketplace for apps running on home routers. It would change the Internet as we know it. Developers should be able to create router apps and submit them to marketplaces and users should be able to install router apps of their choice with just a few taps. Routers have a few key advantages over most other computing devices owned by the public: routers normally have a public (non-NAT) IP address, they…

I think this is a very very interesting idea. The only problem that makes me nervous is telecoms/ISPs capping the line or requiring some propietary interface to it (something done this days to an extent) and quirks. If they feel they are going to lose control or similar. Just a thought, I don't really know if it could be a real problem though.

Re: Re-decentralizing the Web, for good this time

#192

Post by mbrock »

xipho wrote: post #128
tmcw wrote: post #79

It puzzles me that the linked data future is still discussed, as if we didn't already try it, and didn't already discover that developers dislike arcane RDF standards and the academic-rooted designers of the specifications have a terrible track record of solving real-world problems. And that now they're presenting linked data as some critical component of the decentralized web while skipping out on the debates that e…

You're right, there are a lot of academics that like the idea of a semantic web, it rings true to a lot of scientific principles. There are also a lot of ideas and far fewer day-to-day applications. Science ticks a long a lot slower than startup culture, however, so it's not suprising that a) understanding of the issues comes slower, but also b) some "experiments" that would utilize semantic data have not yet been fu…

I’ve been learning about Barry Smith’s project for a scientific base ontology, Basic Formal Ontology, and it’s really fascinating stuff.

Re: Re-decentralizing the Web, for good this time

#193

Post by mbrock »

hinkley wrote: post #189
AnIdiotOnTheNet wrote: post #164

Earlier quoted context omitted.

> I'd argue that it's the practice and not the principle of auto-updates which is unsound. I disagree. I've yet to encounter a system where automatic updates didn't sometimes break things. It's a nice theory, but in practice things break when you change things they depend on even when you're trying not to, and a lot of developers don't try that hard. Not to mention, sometimes you're dealing with design flaws that cau…

I don’t think it would be that hard to build an update system where not everyone gets updates at once, and you wait for consensus among early adopters before forcing it on the remaining users.

You still need a company or foundation stable and competent enough to maintain it and not ruin it for 50 years...

Re: Re-decentralizing the Web, for good this time

#194

Post by zbentley »

cookiecaper wrote: post #125
svachalek wrote: post #103

Earlier quoted context omitted.

I really, really sympathize with the goals here but when I read through these proposals a few months ago I literally facepalmed. They seem about as realistic as praying for some kind of deus ex machina. Ultimately I think there are technical solutions to making the decentralized web more attractive than the walled gardens, but at this point they will need to be ridiculously polished and shiny to even get a look, and…

The internet is already decentralized. Some billionaire can't do anything to fix the situation, at least not directly, because our draconian copyright and network access laws are the only reason that walled gardens are able to exist. The internet doesn't really tolerate serious technical barriers stopping someone from automatically multiplexing the content from various social networks into a single read-write stream,…

> The internet doesn't really tolerate serious technical barriers stopping someone from automatically multiplexing the content from various social networks into a single read-write stream

If the legal barrier went away, and someone surmounted the "unserious" (which I doubt) technical barriers to doing that, then the content providers would go out of business. That's arguably a good thing, but I suspect many people would disagree. The problem is the profit motive and financing model for what consumers want, not the degree of decentralization. Google didn't screw up the internet; people did, by preferring what Google offers.

Re: Re-decentralizing the Web, for good this time

#195

Post by sascha_sl »

zozbot123 wrote: post #86
sascha_sl wrote: post #83

The W3C has a proven track record to produce overengineered shit when it comes to "the semantic web". Just look at ActivityPub. It's essentially OStatus but instead of XML we slapped namespaces on JSON, wrote a bunch of overly complex preprocessing procedures so that everyone can output just the way they want[1] and still made half the spec ambiguous enough[2] that implements essentially follow the one rule that matt…

The Semantic Web does not require XML these days. JSON-LD (for JSON) and GRDDL/Microdata (for HTML) are widely-acknowledged standards. For simple text use, akin to a Markdown-formatted document, you can use Turtle. If you believe that JSON-LD is genuinely ambiguous, take it to the authors of that spec and contribute to getting it fixed.

Do you think a format that requires this amount of code to predictably parse is a good format?

https://github.com/kazarena/json-gold/blob/master/ld/api_nor...

Re: Re-decentralizing the Web, for good this time

#196

Post by zbentley »

kodablah wrote: post #8

I think the part ignored by so many is the need to decentralize the computers into the home. I'm not talking meshes or shared resources. For the majority of use cases, we don't need distributed storage, compute, etc. Just start making these self-hosted "servers", "data pods", etc as easy to install as desktop software and make it clear that they are inaccessible when the computer is off. People that aren't already wi…

> People that aren't already will gravitate towards at least one always-on machine in their house.

I think you radically overestimate the desire and ability of the average computer-user to consider their devices' uptime.

Re: Re-decentralizing the Web, for good this time

#197

Post by zbentley »

eitland wrote: post #181
deevolution wrote: post #171

Earlier quoted context omitted.

Sick, now turn this into a product that can be adopted by the masses!

I have an older Synology. It is already ready for the masses. Why: You only need one one for a family and most of the time there's already a person in the family who does "PC stuff". And even if there isn't there's always someone who'll learn it if a friend has one. The rest of this post is not targeted at you but rather on a whole attitude here at HN: ------------------------------- Anyone who can operate a web brow…

> most of the time there's already a person in the family who does "PC stuff". And even if there isn't there's always someone who'll learn it if a friend has one.

That's not true at all. Confirmation bias is rough when you're technical; you keep spotting other technical people.

Re: Re-decentralizing the Web, for good this time

#198

Post by cjslep »

Karrot_Kream wrote: post #180
ericflo wrote: post #172

Earlier quoted context omitted.

I banged my head against this issue quite a bit last year. Specifically the thing that tripped me up, is how many fields can either have single value or be a list of values, and figuring out what is meant semantically when it's a list or a single value.

This is exactly what frustrates me the most and what most of my logic checks. I'm writing tests for each of these semantic cases, but it seems silly to me that a standard leaves something so ambiguous.

This is why the go-fed project uses code generation.

Re: Re-decentralizing the Web, for good this time

#199

Post by zbentley »

hathawsh wrote: post #35
kodablah wrote: post #8

I think the part ignored by so many is the need to decentralize the computers into the home. I'm not talking meshes or shared resources. For the majority of use cases, we don't need distributed storage, compute, etc. Just start making these self-hosted "servers", "data pods", etc as easy to install as desktop software and make it clear that they are inaccessible when the computer is off. People that aren't already wi…

There should be a marketplace for apps running on home routers. It would change the Internet as we know it. Developers should be able to create router apps and submit them to marketplaces and users should be able to install router apps of their choice with just a few taps. Routers have a few key advantages over most other computing devices owned by the public: routers normally have a public (non-NAT) IP address, they…

Please don't encourage router-makers to add more functionality that risks exposing my network traffic to the internet. I sympathize with the sentiment, but the people that make these devices have proven again and again that they are unable to execute on basic traffic security.

> routers that host arbitrary apps

Servers. You are describing servers.

Edit: upon reflection, rather a lot of people I know don't have routers, either; they use shared internet (e.g. xfinity) or only have mobile plans.

Re: Re-decentralizing the Web, for good this time

#200

Post by maccam94 »

TheOperator wrote: post #101
thsealienbstrds wrote: post #62

Earlier quoted context omitted.

I don't know, if people could buy a 'personal cloud box' that is as user friendly as an iphone, for less than a 100 dollars, a substantial number of them might possibly do that.

The issue such devices have in practice is upfront cost, compatibility, the need for port forwarding, and the lack of redundancy so you don't lose your data. Among other problems but these come to mind. Centralization solves those problems because you can connect to a server instead of port forwarding, your data might be stored across 3 servers, the cloud storage might be as cheap as free, and applications are design…

IPv6 will help by eliminating NAT (and thus port forwarding problems). IPFS might help with redundancy. But yeah the cost and complexity are major problems. I think the biggest problem is making open standards that can evolve quickly. All the big chat platforms have switched to proprietary protocols so they can iterate and roll out changes on their own.
Post reply on HN