[johnnydoebk]

There are a lot of efforts to "re-decentralize the web" and none of them seem successful. In my opinion, that's because their authors are trying to implement some elves fantasies.

1. Has anyone of the "redecentralizers" ever thought about how developers would monetize their software? The web as we know it is popular because there is financial incentive. Current web makes it possible to implement any monetization model out of existing 12: product, service, Subscription, Resale, Audience Aggregation, etc. Why would you want to limit that?

2. Has anyone considered compatibility with the existing applications? Yeah, that's non-trivial, but wouldn't it be cool if with minimal effort we could move existing software to a more decentralized model?

3. Don't tell me which stack of technology to use. Not everybody wants to build JS SPAs.

4. Has anyone ever built a useful and successful decentralized service. I mean the one that would be popular even if it's centralized, not just like "Facebook but federated". How about starting with this step instead of creating an abstract protocol?

[mynewtb]

The web came to be because people could freely express themselves, share their hobbys, interests and works with a global audience. Monetisation played a minor role for a very very long time. I miss that old web.

[nsgi]

That web still exists: look at Wikipedia, one of the largest websites, run entirely not-for-profit with the aim of sharing the sum of human knowledge. Nothing stops people from starting not-for-profit websites today and only linking to other not-for-profit websites. Running a website has never been cheaper and search engines are light years ahead of anything that existed then.

They may never be as popular as Facebook and Twitter, but that's because the web has grown in audience to people looking for something similar to watching TV, reading newspapers and going to the pub. Those activities will always be more popular than people hacking in a garage, but they can coexist.

[hinkley]

The changing audience isn't the biggest reason they're more popular. When people talk about 'stickiness' they're talking about using Dark Patterns of UI and/or using hyperbolic tone (formerly known as Yellow Journalism) to pander or incite, which makes you stick around long after you got the information you thought you came for.

Relatively wholesome sites like Wikipedia can't compete on numbers and hours of visitors because they limit themselves mostly to 'see also', and commentary is split onto another page so you have to choose to see it every time.

[nsgi]

At the end of the day it's peoples' choice whether they choose to read sites like that, which pages they like on Facebook, etc. Those sites are popular for the same reason newspapers like the New York Post and the Daily Mail are popular. For as long as newspapers have been published online, those sites have been part of the web. Maybe not BuzzFeed, but sites posting sensationalist content.

Wikipedia is definitely able to compete. It is the 5th most popular website, behind Google and Facebook but ahead of every online tabloid and content farm.

[RRRA]

They CAN, sure, but they MUST coexist...

and that is less guaranteed considering market forces and network effects.

[qznc]

Has this old web really disappeared or us it just dwarfed by the new web today?

[tqkxzugoaupvwqr]

The old web is still there but it is definitely dwarfed by the new web. Search results are gamed, comment sections are spammed, product placement and paid reviews are rampant, fake neutral information websites erode trust. Because there is money involved and access to the Internet is not controlled, some people chasing the dollars ruin the experience for everyone. They keep generating noise in their chase of money so that eventually the old web is impossible to find because most links lead to the new web.

[qznc]

It sounds like an interesting technical problem at first. A search engine for hobbyists. Use AI or whatever to filter out commercial interests.

The problem is that nobody will pay you by definition and you are in an arms race against commercial interests. Does not sound sustainable.

[jsilence]

Remove the cause and not the symptoms!

[qznc]

Remove capitalism?

[reitanqild]

Parts of it still exists I think.

I'm afraid we help in killing it everytime we:

- Complain about styling and form

- Worry about the SEO effect of what we link to

- etc

I wish people would start to create more weird and wonderful pages. Pages without styling. Pages with nice styling. Pages with weird styling. But most importantly full of text content, links to friends and weird projects we love, images and videos etc.

[mr_spothawk]

dwarfed, ya. and it appears that the "new web" participants engage in legal attacks on, e.g. popular blogs about hacking owned hardware. so, dwarfed and maybe bullied, too.

[tuxt]

Someone has to pay the bill.

[a3_nm]

If you want to write things on a self-hosted static blog in your free time, the hosting costs add up to a few dollars per month for a domain name and (e.g.) a VPS, for something that can scale to thousands of simultaneous readers. This is completely negligible, so I don't think "the bill" is a very important factor.

[astrobe_]

We already do. The hard disks, the bandwidth, the electricity.

Give me a ready-made virtual machine image [1] pre-configured to share files. I don't care about the protocols, I don't care about super-duper-unbreakable encryption because I won't share or host copyrighted content nor wombat porn.

[1] Not a freaking docker image or whatnot (hi, Mediagoblin!) - if you want it to succeed it has to be dead simple so that mere mortals can do it.

[calgoo]

So my idea would be to build a private cloud for your house. The idea goes something like this:

Sell a "box" (think something that looks like a apple airport extreme). The device will have a nfc / bluetooth / wireless / rj45 and internal storage. To link your device to the private cloud box, you use the app on your device and sync public encryption keys. Now there would be two ways of syncing photos / data between the cloud box and your device:

1. Local Wifi sync - when the device detects that is on its home network it will sync any changes.

2. Remote Sync - Use a public server, upload encrypted data using the public encryption keys, and the home cloud will download them.

For sharing with others, im thinking of of web / circle of trust, where you can share your public keys with your friends / family. There would be privacy levels: 1 - Me & wife / 2 - Close family / 3 - Close friends etc...

Anyway, this is just an idea i have been thinking about for some time now. There is still a lot of details to work out but I do believe that we need to try moving away from the public cloud.

[tyingq]

It would be interesting if it were possible to hijack the infrastructure of the existing cloud players to do something similar. Like using encrypted blobs on Google drive, Microsoft OneDrive, etc, as the backing store. In some way that provides redundancy, safe peer-to-peer key exchange, opacity (your account is storing not just your encrypted blobs) etc.

Foisting off the infrastructure funding on the opposition has a sort of poetic justice to it.

[nickpsecurity]

Google "parasitic storage." That's what CompSci called such schemes when I last saw them. They might even have done "parasitic computing" or something like that by now.

[tyingq]

Interesting. I could see using something like the blockchain as proof of which domains/urls are the current central authority to get the rest of the info from. Such that the base reference isn't easily shut down. It's whack-a-mole to try and kill the head, no need for complex decentralization.

Would be kind of fun to watch the giants trying to shut it down. Given some level of popularity, it might become too difficult​ for them to do. Users donate newly minted Google/Microsoft accounts as old ones get shut down. With some sort of karma reward for doing that.

[jude-]

You might want to take a look at Blockstack (https://blockstack.org/). It does exactly this on top of the Bitcoin blockchain. You resolve a globally-unique name to a DNS zone file, which has URI resource records to off-chain data on one or more external storage systems.

Disclaimer: I'm one of its creators.

[tyingq]

Only on HN. I describe an unusual idea in fairly vague terms, and it turns out it exists. And one of the creators clues me in.

Thank you, and I will check it out.

[qznc]

Use an old laptop as box.

Install NextCloud, eg via "snap install nextcloud" on Ubuntu. Now you probably want to configure NextCloud a lot via its web interface.

You are 90% there. There might be a hurdle to access your box if it is behind your home NAT.

Not available to non-techies, but you could sell such pre-configured boxes.

[rodolphoarruda]

I would rather use my phone as the box instead, since it is the device that is potentially on 24/7.

I wish that by installing the app it would take me to the following process:

1. Set up my custom domain preferences 2. Set up email 3. Set up file sharing 4. Set up personal website (optional)

For the periods in which my phone is offline, e.g. on a flight, the whole thing would fall back to the app vendor service as a contingency to people sending me email or wanting to access my website. Once I'm back online, things would be sync up to the phone and the fall back contingency removed.

Maybe it's just me wanting to fast forward the world 5 to 10 years from now.

[qznc]

A phone is not "on" 24/7. A phone is asleep 99% of the time, because anything else would drain the battery too quickly.

[educar]

Why restrict the software to nextcloud? What if I want Git hosting? Or run some cms. Imo, something like yunohost, arkos or cloudron is better suited.

[ocdtrekkie]

This, one of the key perks of Sandstorm.io is that you can use apps written in any language that runs on Linux pretty much.

Cloudron.io is fairly similar, though makes some different design choices.

[educar]

I definitely like sandstorm approach over nextcloud plugin model that tries to make nextcloud do everything. It's a security nightmare

[ocdtrekkie]

I do too, and I feel Sandstorm has the superior security model of most of these platforms. But a key perk of Nextcloud is that it can be installed on cheap shared hosting which doesn't grant the customer root access.

[qznc]

Sure.

Still, NextCloud provides the most popular cloud stuff for non-techies: Storage like Dropbox and adressbook like Google/Apple.

[newsat13]

Do you know how well NextCloud works in a home environment? Is it really easy to automate the punching of a hole in the home firewall to allow access from anywhere? If you know of a project that automates this punching, it's a great combo to nextcloud.

[qznc]

I punched the hole manually. Unfortunately, even this is not enough. The tricky part is to have the same hostname inside and outside of your NAT.

[retor]

Have you tried a Diskstation?

[tyingq]

>4. Has anyone ever built a useful and successful decentralized service.

It does seem like "incentive" is the missing piece. The two somewhat successful, somewhat decentralized things I can think of are Bitcoin and TOR. Both provide some compelling incentive to either join as a user or to subsidize infrastructure.

[m-j-fox]

And BitTorrent.

[nickpsecurity]

"4. Has anyone ever built a useful and successful decentralized service. I mean the one that would be popular even if it's centralized, not just like "Facebook but federated". How about starting with this step instead of creating an abstract protocol?"

BitTorrent is the only one that I'm aware of past the core Internet protocols. As I said elsewhere, it spread because of the problems it solved for users in a way they could learn easily. Most probably didn't even know it was decentralized. So, best advice is to solve a problem for users really well with extra benefit of "we also protect your privacy instead of selling you out."

[pherq]

Usenet, back in its heyday, might also qualify

[nickpsecurity]

Oh yeah. It did get a lot of usd at least in the early days.

[PuercoPop]

> Has anyone ever built a useful and successful decentralized service.

E-mail fits all of your requirements. Even the now being centralized (Gmail)

[syrrim]

>2.

I have thought fairly hard about compatibility. What you could do would be to serve a standard website over http, but allow anyone to act as a cdn. The only centralized piece would be dns to route people to their closest peer. Most clients would think the peer was a normal http server, under control of the website owner. In reality it would be any random person on the internet who had decided to rehost the content. Therein lies the problem: how do clients ensure the integrity of resources if they don't even know to look for it. One of your cdns could be serving fraudulent content, and none of the clients would question it.

In order to distribute http, you would therefore be forced to get at least some custom code into browsers. At that point, why not shove a whole ipfs implementation in?

[syrrim]

>4.

Yes. Bittorrent. It makes use of centralized directories, and centralized trackers, but is otherwise p2p. It suffers from everything weve come to expect from decentralized services. It is only able to survive because no startup can legally disrupt it (not to say no one's tried). In order to use it, a person first has to download a custom client. They also have to visit a completely separate website to get a list of possible torrents. Both the most popular client and the most popular website are riddled with obscene and intrusive ads. What it should have, if it wanted to compete with netflix and spotify, is a web client with the ability to play things instantly. Both those things are antithetical to bittorrent, and so it will continue to lose to the legal avenues.

[Sir_Cmpwn]

Mastodon/GNUSocial seems to be the counter-argument to most of your points.

[shea256]

Hi! This post really resonated with me because of my experience with how previous efforts at re-decentralizing the web have failed.

In my view, adoption will always be held back as long as there's:

a. a UX gap (in both software and dev skills)

b. a lack of innovative open source business models

c. an absence of completely novel apps that couldn't have been done before

We're trying to tackle all 3 at Blockstack with our developer platform but it's not easy. Would love your thoughts sometime on whether we're on the right track and how else we could help you out as a developer.

[laughingman2]

> The web as we know it is popular because there is financial incentive.

Really? how does wikipedia function then, why do lots of people create meaningful content in web for free?

Applying the results of a system (aka web being currently commercialized) as its cause is circular reasoning at its finest.

[curt15]

>Really? how does wikipedia function then, why do lots of people create meaningful content in web for free?

I get the impression that sometimes it's barely getting by; very few months I see a banner at the top of its webpage suggesting that it's close to running out of money and pleading for donations.

[dorfsmay]

Aren't the different static generators a good step towards re-decentralising the web?