Neocities is implementing IPFS – distributed, permanent web • Hacker News Board

Re: Neocities is implementing IPFS – distributed, permanent web

#211

Post by smacktoward » Wed, Sep 09, 2015, 2:16 PM UTC

enkiv2 wrote: post #128

Just to point out -- hypertext systems prior to the web almost universally expected permanent stable addresses (and this was, for instance, required by the various Xanadu specs). Enquire (Tim Berners-Lee's immediate predecessor to the Web) had permanent stable addresses, too. The web didn't have permanent stable addresses because the web was a hack, intended as a way to explain Enquire to people who didn't have a bac…

This turned out to be not a bug, but a feature: removing the requirement for permanent, stable addresses made a large-scale hypertext system something that could actually be implemented. (Unlike Xanadu.)

Here we are, more than twenty years later, and a hypertext system with permanent, stable addresses is still so rare in the real world that creating one will get you on the front page of Hacker News. The "crappy" Web, meanwhile, is the foundation of modern commerce, culture and communications.

If that's crappy, I hope to God I can make something crappy someday.

Re: Neocities is implementing IPFS – distributed, permanent web

#212

Post by nightpool » Wed, Sep 09, 2015, 2:41 PM UTC

sdoering wrote: post #79

_prometheus wrote: post #20

Earlier quoted context omitted.

Two pieces of (coming soon) good news: - very soon, you wont need to install anything to use IPFS. it will "just work" with js on today's browsers. - for best perf, yes, we need browser implementations. and... those have begun :)

OK, as apparently .js is the defacto thing to build cool sites, I for one will be left out here. On sites not known to me, I typically block most (read all) .js at first, just to make sure that these sites do not leak too much information about me to external tracking tools, advertisers or the like. So using .js seems to me a bit problematic in regards of people not enabling it by default (but this is an absolut mino…

The first implementation of IPFS is written in Go, and it includes fully functional nodes, http gateways, and libraries, but they're also working on a node.js version (that's also compatible with browser environments) so that non-technical users can have ipfs sites "just work". (although it requires at least static content served from a "normal" http/html server)

Re: Neocities is implementing IPFS – distributed, permanent web

#213

Post by mileswu » Wed, Sep 09, 2015, 2:41 PM UTC

njloof wrote: post #208

rdancer wrote: post #198

When a massively distributed file system[1] was introduced into the Linux kernel circa Y2K, there was what I remember as a lot of hype. Nobody, but nobody wanted it then (or since, it's still part of the kernel), and I would be surprised if anybody wanted something similar today. [1] https://en.wikipedia.org/wiki/Andrew_File_System

All undergraduate accounts at Carnegie Mellon were on AFS in the 1990s. It may have been ahead of its time -- all I know is (a) it was much less reliable than NFS; (b) its security policy was different enough that people (admins included) left lots of security holes; and (c) *.struck

CERN still uses AFS extensively for all users' files. I believe they use a forked version of OpenAFS. My experience with it at CERN has been okay. It's been fine to use locally, but transatlantic over the Internet can be very slow (it does work eventually though).

I'm not sure if there are any alternatives that work any better over long distance. I suppose CERN also has CVMFS [1], which is a read-only caching filesystem that retrieves the files from upstream via HTTP. This works much better, but it is read-only so only satisfies certain use cases.

[1] http://cernvm.cern.ch/portal/filesystem

Re: Neocities is implementing IPFS – distributed, permanent web

#214

Post by hardwaresofton » Wed, Sep 09, 2015, 2:45 PM UTC

walrus wrote: post #201

hardwaresofton wrote: post #47

Earlier quoted context omitted.

Right, but fiddling with probabilities is exactly what you want to do in that case right? I don't know if a hashing algorithm that has no collisions exists (I suspect the answer is no, maybe other than the identity function). If collisions start to become a problem, you can move that problem further out (much much farther out) by mixing more hashing functions in, which was my point. It's well documented (of course al…

> I don't know if a hashing algorithm that has no collisions exists (I suspect the answer is no, maybe other than the identity function). There is no hash function without collisions. The set of inputs is infinite but the set of outputs is finite. (The identity function isn't a hash function--its output isn't a fixed length.)

Thanks for the correction! Somehow forgot the fixed length output requirement of hashing functions.

Re: Neocities is implementing IPFS – distributed, permanent web

#215

Post by zkhalique » Wed, Sep 09, 2015, 2:48 PM UTC

ZenPsycho wrote: post #92

zkhalique wrote: post #84

We've been thinking about this problem for quite some time. My friends and I are passionate about re-decentralizing the web. The biggest challenge we've come across is this: Many users want to do things with other users That means there has to be a way to persist state and history (of a group activity, an evolving resource, etc.) across machines. Mental poker is still very hard to do for groups. Since most clients wi…

Couldn't you do group editing through using lisp inspired link lists? See a comment thread? CONS your comment onto the end of the comment thread and push into the ether, and notify about the new list head. Take other patterns from clojure immutable data structures for various other kinds of group edited time travelling sites

It might be possible to do some sort of self contained data structure that authenticates everybody and only allows authorized access to it, which solves the Byzantine Generals problem like BitCoin does, and which resists forking and instead uses vector clocks and Operational Transformations to consolidate actions done to the tree. But even so, how would it resolve conflicts? It could be a long time until a conflict is discovered, and then do edits get dropped, or what?

Re: Neocities is implementing IPFS – distributed, permanent web

#216

Post by zero_iq » Wed, Sep 09, 2015, 2:48 PM UTC

radlad wrote: post #22

jeswin wrote: post #21

Earlier quoted context omitted.

Since the private data and the query are from the same user, they could share some mathematical connection perhaps, which would allow a query to be done and the results sent back (results which only the user can decrypt). Google shows up plenty of articles on the subject, but I am not an expert to know where we stand on this.

Interesting. So you'd perform operations on encrypted data, and send back an encrypted result that's useful yet different from the original encrypted data, all without ever performing decryption? It seems possible but extremely tricky.

https://en.wikipedia.org/wiki/Homomorphic_encryption

My understanding is that it is possible, and Turing-complete systems can be built, but they are slow and unwieldy, requiring a large amount of cyphertext to encode data (on the order of 4KB per float, I seem to recall).

It's still up for debate whether or not such techniques will eventually be up for general distributed computing, e.g. distributing a VM instance over multiple distributed servers; or just for specific use-cases, e.g. bitcoin-esque distribution of small transactions, etc.

Re: Neocities is implementing IPFS – distributed, permanent web

#217

Post by kordless » Wed, Sep 09, 2015, 2:48 PM UTC

rdancer wrote: post #198

When a massively distributed file system[1] was introduced into the Linux kernel circa Y2K, there was what I remember as a lot of hype. Nobody, but nobody wanted it then (or since, it's still part of the kernel), and I would be surprised if anybody wanted something similar today. [1] https://en.wikipedia.org/wiki/Andrew_File_System

> uses a set of trusted servers

This uses a set of untrusted servers, so it's different. Also, trust use-cases come in a lot of different flavors, so the comparison of 'wants' is pretty thin.

Re: Neocities is implementing IPFS – distributed, permanent web

#218

Post by nightpool » Wed, Sep 09, 2015, 2:49 PM UTC

kang wrote: post #195

reit wrote: post #191

Earlier quoted context omitted.

>In case of neocities, a central server is giving you a hash that the nodes would provide you the data for, but without a central management how does IPFS associate correct file to the correct hash? Mathematics[1]. Your hash is based on the content of the file, therefore you can guarantee that the content is the same as what the original author published. Also, IPFS is not a blockchain, there is no 51% attack. https:…

I am not native english speaker so perhaps was misunderstood again. You searched for a content by an author, and the network gives you multiple hashes, which one do you trust, without a central authority telling you?

I think the fundamental misunderstanding here is that you can't search by author. You can fetch by IPFS hash or IPNS pubkey (which is tied to a hash). You can also use existing name services (such as DNS or namecoin) to tie human-meaningful names to a pubkey or hash. (but probably a pubkey)

Re: Neocities is implementing IPFS – distributed, permanent web

#219

Post by kordless » Wed, Sep 09, 2015, 2:52 PM UTC

nly wrote: post #174

A universal distributed file system that looks native on my system (thanks to FUSE) and caches content I access locally, while serving it securely to others, is interesting... assuming you can solve key problems like search, identity when you want it, privacy when you dont, and human memorable addressing... which afaict IPFS doesn't address. The "permanent web" on the other hand you can create using MHT archive files…

> Wake me up when someone figures out how to make apps that require seamless multi-user collaboration, like say distributed clones of HN, Facebook or Github.

There are people working on this, so stay awake.

Re: Neocities is implementing IPFS – distributed, permanent web

#220

Post by treve » Wed, Sep 09, 2015, 2:53 PM UTC

enkiv2 wrote: post #128

Just to point out -- hypertext systems prior to the web almost universally expected permanent stable addresses (and this was, for instance, required by the various Xanadu specs). Enquire (Tim Berners-Lee's immediate predecessor to the Web) had permanent stable addresses, too. The web didn't have permanent stable addresses because the web was a hack, intended as a way to explain Enquire to people who didn't have a bac…

Stable links were also part of the earliest ideas for the web, they just failed early. The URL was only intended as the 'locator', the URN was the stable link, and some way had to be found for URN's to be mapped to URL.