/ipfs/ (old thread http://boards.4chan.org/g/thread/51 45972

Gateway link gives a 404, fyi.

>>51467723
i bet you didn't give it a hash

>>51467723
Works fine here. What file are you accessing?

is there a /g/ipfs irc channel? or should I make one? I know they have one but I mean one on rizon for /g/tard to share files

IPFS whitepaper draft: https://gateway.ipfs.io/ipfs/QmXAQdSkFbJHCPHDaALSxe qtT2fFFV3U1FxeSvDmXVFgUD/ipfs-draft3.pdf

>>51467748
registered #/g/ipfs on rizon

>>51467723
https://ipfs.io/ipfs/QmSd1buim52M tQP1SH2XhN7FC8c1 KdFrFKBNmgdUKFdss8
(remove spaces)

the talk linked from the OP, i grabbed it from youtube and am seeding it from a vps with a high speed internet connection

how fast can people download it? i'm curious (post your results!) (preferably using a local ipfs daemon)

More music
Qma5FxHL5rTo1EUmbG
Tzv5K18BPzzzgtYAy45hzzXX3hGL

>>51467883
It went up to 1.40 MB/s, then suddenly stopped.

>>51467953
hold up, i think the binary i put on the vps was compiled for the wrong cpu

>>51467910
Thanks mate

[file] Reading from stdin...
 (+) Audio --aid=1 (flac)
File tags:
 Artist: The Bloody Beetroots feat. Tai Jason & Bart B More
 Album: Hide
 Album_Artist: The Bloody Beetroots
 Genre: Electro House
 Title: Spank
 Track: 1

>>51467953
>>51467974
try now

File: nofun.png (23KB, 566x149px) Image search: [Google]

23KB, 566x149px

>>51468057
Still does not continue downloading

>>51468057
I'm on 52%, and it's still downloading, but it's incredibly slow (~150kB/s).

>>51468187
It's fast enough to watch it in real time, though.

Explain briefly: What is the big deal with this IPFS meme?

>>51468508
/g/ thinks everyone will want to keep a static copy of popular websites in their hard drive and leave their computers always on as a server for backups to create a distributed HTTP

>>51468540
>everyone
No. And you don't have to pin it to seed it. Anyone using the network will share it until it's garbage collected.

>>51468553
You don't? What are you seeding then?

>>51468565
You seed both files you have accessed recently and those you have pinned.

>>51468576
...how does that contradict my original post, then?

You access ipfsgoogle
you become a "backup server" of a sort (it's actually worse) for ipfsgoogle, which means some of your bandwith will always be taken if ipfsgoogle is a popular website

>>51468588
I pressed enter too fast. It's TOR's problem. Early adopters will have to suffer more than if the system was already widespread. Which means there'll be less adopters, and the system will never be widespread.

The only way this might happen is if google pushes it, but they're the US government's bitches and they won't back a system that takes power away from them.

>>51468588
>>51468604
>...how does that contradict my original post, then?
You're implying everyone must be an active participant in the archival process for this to work. They won't.

>Early adopters will have to suffer more than if the system was already widespread
I'm not sure how true this is. The ratio of providers to users should remain more or less constant. Have you done any calculations or are you just speculating?

>>51468715
Now that I think about it, you might be right. I didn't think about the ratio. Of course, we're basically using a more obscure http right now since there aren't enough users to make a visible difference, but you've convinced me the idea has growth potential.

And everyone will be an active participant in at least ONE archive. The moment they access a website they become a seeder. I have a jewish ISP that restricts seeding, and I'm european so the situation in the US is probably worse, so people might not want to waste their limited seeding bandwith in redistributing some parts of an anime video file to users.

But again, you've convinced me anyway. They're working on some sort of DNS for this, right? Making hashes readable?

>>51468764
>we're basically using a more obscure http right now since there aren't enough users to make a visible difference
True, but still: Think about how few peers are actually needed to get decent speeds for a torrent. Even 10 seeders usually caps out my download bandwidth. The same thing should apply here.

>people might not want to waste their limited seeding bandwith in redistributing some parts of an anime video file to users
You have a point. It would be really sad if this failed because of ISP fuckery.

>They're working on some sort of DNS for this, right?
Yes. For now, you can leverage the normal DNS system to insert a link between your regular domain name and an IPFS or IPNS address. See: https://github.com/ipfs/examples/tree/master/examples/websites
There have also been talks about adding support for Namecoin for address resolution, which would make every part of the system distributed. I'm not convinced that would take off however.

>>51468847
You seem to know a lot about ipfs, have there been any tries at a distributed HTTP before? It's weird that I haven't seen any. I know about CHORD, but as far as I know that was just trying to be p2p and just a concept.

it's pretty interesting what you can find just by using scraped gateway addresses on a search engine

someones' old computers
http://localhost:8080/ipns/ipfs.bennyp.org/nerd/

some usb hunt thing
http://localhost:8080/ipfs/QmSJu9Mw
nccFAuau1CFsWQRxVat
BK5nVdXHq4tvrJ6iE1H

some junk
http://localhost:8080/ipfs/QmTnV2DBMwTsvj8Pga7
a2rYruvhrXVbVxob
smTEYBkYTrw/Pant%20Zipper

ipfs archives
http://localhost:8080/ipfs/QmZBuTfLH1LLi4JqgutzB
dwSYS5ybrkzt
nyWAfRBP729WB/archives/index.html

interplanetary wiki
http://localhost:8080/ipfs/QmdDbLyNKCwt
BVFhEVQfdBTCnEWmW
RJGnm93qa3kzBjVym/#!/home

>>51469004
whole bunch of hentai manga
http://localhost:8080/ipfs/QmchsYMr8Xc3y52Ab
Sa5QAqJSWqwKTp
oeZ7JpBSTw5xMCR

>>51469043
whole bunch of turkish ebooks
http://localhost:8080/ipfs/QmUC4KA1Vi3DizRr
Tj9Z4uyrL6a7zjS7
wNnvR5iNzYALSh

>>51468977
The original Project Xanadu has a lot of similarities with IPFS, at least on a conceptual level, but it predates HTTP by a few decades, obviously. Other than that, all other systems I know of (9P, AFS, Tahoe LAFS) have mostly been intended for p2p file-sharing, like you say, or as "regular" systems of files and not a drop-in replacement compatible with the existing web.

>You seem to know a lot about ipfs
I've been playing around with it a bit for the past couple of months, but nothing more.

>>51469004
>scraped gateway addresses on a search engine
explain

>if you find a vulnerability that may affect live deployments -- for example, by exposing a remote execution exploit -- please send your report privately to [email protected]. Please DO NOT file a public issue

why they don't want it to be public ? It's still alpha software, why not keep public aware of vulnerabilities and make them know which are found, need to be fixed and are fixed ?

>>51469220
r u retarded

>>51469129
as in, on google or bing or w/e, search for;

site:ipfs.io/ipfs

or

site:ipfs.io/ipns

File: chuuniwoop.gif (1MB, 640x600px) Image search: [Google]

1MB, 640x600px

>>51469004
>http://localhost:8080/ipfs/QmTnV2DBMwTsvj8Pga7
>a2rYruvhrXVbVxob
>smTEYBkYTrw/Pant%20Zipper
Heh, that's my archive that I shared with someone from /f/.

>>51469256
your shame is permanent now

>>51469236
Found a whole bunch of what seems to be Creative Commons licensed Arxiv papers: http://127.0.0.1:8080/ipfs/QmfXH9XtP7xmoTH8WAp4HNSduqWMwLTH8B8TvbTkdgzNAa

>>51469301
that's actually linked on >>51469004
"ipfs archives" as well

>>51467711
>BitTorrent on steroids
does this mean it requires trackers or other servers that connects other peers?

>>51469442
Nope, it uses DHT, similar to how magnet links work.

>>51469442
no, there are no servers, only nodes

>>51467711
Can someone tl;dr what ipfs is ? I'm gonna watch the few videos presenting it, but would like to know what it is about before.
For what I understood it's a protocol for a decentralized web encrypted ?

>>51469466
watch the videos, any tldr will effectively be a shitty transcript of one of the videos

>>51469466
If you just want to know what it is, basically bittorrent for HTTP. You get a part of a website from every user. The twist is that it's completely decentralized and actually (trying to be) web-oriented.

Someone expressed they wanted to make an irc channel, please do.

>>51467723
The gateway isn't meant to be used directly. You have to do gateway.ipfs.io/{ipfs|ipns}/<some-hash>

I wanted to add the files posted in the last thread but I can't into metadata. Please post and give a name, or use -w.

>>51469479
it goes a step further than decentralized, it's distributed

decentralized means anyone can be a server
distributed means everyone IS a server

>>51467883
Jesus christ IPFS is fast, that 339MB webm loaded fucking instantly.

>>51469514
of course it is when bunch of /g/ people are connected with fucking gbit links

>>51469493
yeah I messed up

>>51468764
>I have a jewish ISP that restricts seeding
Torrent seeding. They can't restrict simple uploads.
This isn't just for sharing files.

>>51469514
the more people check it out, the more times it's cached and the faster it gets

the exact opposite to 'conventional' methods which get slower and slower as more people connect to the same server

>>51469547
>>51469580
Is the file split up between peers?
Like if i request a 1GB file and there are 20 peers connected that have the file would i be getting 50MB from each peer?

>>51469514
I see 13 peers at the moment. Kinda goes to show just how few people are needed to give you insane speeds.

>>51469603
i don't think so

what if one peer is down?

>>51469603
that's is possible

the file is internally split into smaller chunks and chunks can be downloaded from whoever, in this sense it works just like bittorrent

>>51469603
Yes, it's split into blocks, and requesting the file can lead to different blocks being downloaded from different peers. Every peer usually owns the whole file, though (but they can override it).

>>51469574
I know it isn't. And I didn't mean just that kind of restriction, I mean the fact that I literally have 1/10 of my download speed as my upload speed

>>51469479
>>51469493
>>51469563
I understood that but I don't get how
>You get a part of a website from every user. The twist is that it's completely decentralized and actually (trying to be) web-oriented.
is possible. I guess most of my questions will be answered if I watch those videos, but how can you make sure a website can be accessed. Imagine a website of an association that is visited once per week and which nobody really care about. How can you make sure it will be accessed ?
And what about websites like 4chan or any forum ? How can you make sure all the people will have the same updated version ? Or maybe I don't really understand it and I'm just don't get how it works.

>>51469622
then it'll get the piece from someone else who has it

>>51469463
Just curious, but how is a new peer bootstrapped for the DHT to work?

>>51469622
Then it will get it from another peer obviously.

>>51469641
yeah i guess so

>>51469233
The standard way is to report it privately and publish it after a week or so. If you made it public day one the bug could be abused before it's fixed.

>>51469640
>Imagine a website of an association that is visited once per week and which nobody really care about. How can you make sure it will be accessed ?
That association will be hosting the website obviously. If it's not visited they'll be the only host unless someone pinned it for some reason.

>>51469640
In the worst case scenario, with an unpopular and seldom visited site, it will work as a HTTP server, more or less. It still has some added bonuses, like content-addressability, easy mirroring and encryption built into the protocol, but it will be server-to-client like with HTTP.

>>51469639
>I mean the fact that I literally have 1/10 of my download speed as my upload speed
Unless you have a bandwidth cap you're fine. Having lower upload then download is pretty normal unless you have fibre.

Does anyone know how to get a nice ipns name like benny's site?

ipfs.io/ipns/QmaGks9KKzu2
WykHQjJFJkcUAN4Zo
F7ok9h2hXj1WQn47U/

updated with moar stuff. And reminder to install gentoo!

Keep posting links you find, if you're posting a file without a filename please also post the filename. Tell me if I'm missing anything.

>>51469710
All you have to do is set a TXT record on your DNS.

$ dig txt ipfs.bennyp.org

ipfs.bennyp.org.    900    IN    TXT    "dnslink=/ipfs/QmRNChTrdEaFdzsVRrBw68XZpLCSGsW4P1H1efjRPqrAZy"

The format has changed recently, and might not work for IPFS clients. Google "IPFS dnslink" if you want details. The format he uses will work for all future versions, though.

>>51469674
>encryption built into the protocol
[citation needed]

>>51469777
Oh, and you can set "dnslink=/ipns/Qm...", and the resolution will be recursive. This works when looking up the path in a web gateway, but NOT from all command line tools (because of a bug that hasn't been fixed yet).

>>51469777
How do 3rd parties access your DNS's records though?

>>51469810
Via a standard DNS lookup. This is external to the IPFS network.

>>51469784
ipfs.io

>>51469759
Could you post the perma ipfs link? Ipns isn't resolving

>>51469644
Pls respond
I'm curious too

>>51469644
>>51469835
me too, OP explain this shit

>>51468576
How does this prevent abuse?

>hey check out this link anon :p
>hard cp
>you accessed it so ipfs cached it
>youre now seeding cp to pedos

>>51469825
Can't find a single mention of already implemented encryption in that website

>>51469834
QmYLmYbXxnNZ75A
FcCPG4knoqoVZQZG8ZYW
rrTvB7QdVfH

>>51469895
this, it isn't even taken into account in US laws so you'd be fucked even if you weren't guilty

>>51469895
It's encrypted, and you can clear your cache manually. It's not really any worse than having CP in cache because someone posted CP on 4chan over http.

>>51469644
>>51469835
>>51469867
I think you need to include some "bootstrap" peers. You can see them in your config at ~/.ipfs/config. Not sure how these peers are decided upon, though.

IIRC, torrent DHTs also have such a list.

>>51469644
https://github.com/ipfs/examples/tree/master/examples/bootstrap
>IPFS comes with a default list of trusted peers, but you are free to modify the list to suit your needs.
i.e. like every p2p network ever - Freenet, Tor, Bitcoin, i2p all do it like this.

File: 1610.jpg (140KB, 1500x1292px) Image search: [Google]

140KB, 1500x1292px

>>51469925
>nsa servers as default peers

>>51469895
>>51469911
>>51469912
I think there needs to be support for blocklists that you can subscribe to, just to be on the safe side.

Aight, I got a cool name for my sites list page: /ipns/ipfs.gref.org/ will now lead to it. I guess you'll have to wait for DNS propagation first to use it yourself, and I'm not even sure if it'll work, but wekrs for me :^)

>>51469912
for the last time, it's not encrypted, the filename name is just hashed, but anyone can tell which IPs have which file

>>51469969
It's 100% encrypted. Nice FUD though.

>>51469925
does that mean OP have access to all IP addresses that downloads/uploads files over IPFS?

>>51469972
explain how anything I said was wrong, also you didn't reply to >>51469900 you double nigger

>>51469969
Communication is encrypted. You can't see what is being transmitted while it is in transit. But yes, if someone requests it, your client will respond that you own the file.

QmTgNJEgQaCqRht9KSXNyZsCp2xpHZmBRms28NRMmtcERp

>>51469969
>filename name is just hashed
FYI, there are no filename hashes going on. The whole file is hashed to form the IPFS path.

>>51469977
yes
and anonymity would need to be implemented on top of ipfs but they're delaying its implementation until it's barely possible

>>51469991
what you're saying is completely pointless encryption, brah, the mere fact you're going to seed the content you just accessed means you can extrapolate the transmitted file

>>51470007
yeah I fucked up
I wish there were double positive rules so filename name translated to file

>>51469983
"ipfs uses encryption for all communication, but it's NOT PROVEN SECURE YET! It may be totally broken. For now, the code is included to make sure we benchmark our operations with encryption in mind. In the future, there will be an "unsafe" mode for high performance intranet apps. If this is a blocking feature for you, please contact us."
"* filesystem-level encryption"

>reading is hard

>>51470008
Use i2p or tor if you want anonymity with ipfs. This is the correct approach: using an existing, vetted, well-audited solution instead of trusting yet another implementation.

>>51470045
I fail to understand what that encryption is even for if everyone that has the hash can both see what the file is and who has it

>>51470007
What about collisions?

How does ipfs decide what to download/cache/seed? If I go to a couple of the pages in this thread (that also have links to other pages) and then do a "ipfs repo gc", it removes hundreds of hashes. Does ipfs automatically download data from unopened links (if they're linked to by another ipfs page)?

>>51470062
MITM protection for one. Technically, you can't see what the file is until you access it though.

>>51470066
If you could find sha2 collisions you could do more interesting things with them, like stealing all the bitcoins.

>>51470066
>less than 1 in 10^45 chance of collision
>what about collisions
KEK

>>51470066
Then you have bigger problems than IPFS breaking down. (And it's constructed in such a way that it can be patched in the future).

>>51470076
How would MITM attacks even work in a distributed system?

>>51470068
Those are the hashes of blocks forming larger files. Lots of video and music files have been posted, so it's not surprising you have lots of blocks.

>>51470092
Doesn't the creator expect this to grow to become a Galactic-level file system? There would be collisions on such a level.

>>51470068
Pages contain more than one object, namely they contain several chunks per item potentially, and also you'll have the js component, the html, the image, etc. Only what you access gets put into cache. If you pin things, they stay. If 90% of your cache is used, things get gc'd naturally.

>>51469912
>It's not really any worse than having CP in cache because someone posted CP on 4chan over http.
the difference is that in ipfs you are sharing the cached cp with others even if you didn't want to.
which makes me wonder, why isn't there just an "Only seed pinned files" option?

>>51470118
No there wouldn't. You'd need to magically get into contact with thousands of entire alien civilizations immediately for this to even remotely be possible. Nevertheless, the system is implemented in a way that it's fully upgradeable at any time. Hashes aren't sha2 exactly, they're multihash which are currently implemented with sha2.

>>51470136
Because then the system would be too centralized probably.

>>51470144
I see, that's pretty thoughtful of the creator.

>>51469759
>>51469902
I still can't access your files. Is your daemon running and all that?

About CP/anonymity/encryption issues: I'd say that not focusing on those is actually a feature of IPFS. We have all those great networks, Freenet in particular, which does basically the same thing as ipfs, but nobody is using it because it is slow as shit.
Your average normie doesn't need Freenet-level security for downloading pirated movies, cat videos and shitty memes from leddit and that is what ipfs is trying to replace.
If you are a pedo and need pedo-level anonymity, go use Freenet. It works right now.

File: 6288a9f67e09b0165101f996a4d38c78.jpg (310KB, 1200x1200px) Image search: [Google]

310KB, 1200x1200px

>>51470208
It is opening fine both on gateway and on my local daemon

>>51470208
works for me and from ipfs.io. The problem seems to be on your end.

>trying to cut out the jews from the internet
This guy is gonna die soo fast.

>>51470281
don't worry, i have him pinned

>>51470058
so why do we trust ipfs in the first place?

>>51470249
>>51470250
Huh, weird. It's not listing any peers when I run

$ ipfs swarm peers

.

>>51470136
because nobody would seed then

>>51470332
If a file you have pinned remains accessible even when the original uploader has disappeared, then you know you can trust it.

File: 1397836321284.gif (260KB, 266x207px) Image search: [Google]

260KB, 266x207px

>>51470333
Forward port 4001

>>51470333
Try restarting your daemon.

>>51470360
uPnP bruh.

But it turns out my network card was acting up, as usual. I can access it now.

>>51470332
Because you can easily read the code for the transport protocol, and failure is non-critical. The encryption protocol they use needs to be audited more thoroughly but that's part of the "minimum necessary set".

>>51469925
I can see someone doing malware botnets over this shit when its private enought

so.. anyone up for torrent tracker-like index?

>>51470585
Make one and I'll link to it. For now there's my content index at >>51469902
>>51469759

>>51458314
>that being said I do like systems like Perfect Dark and Freenet where niche stuff can survive because of forced distribution. I want it but also don't.
Of those sort of systems, I've only used PD. And I found that it has an interesting rating system in place to get rid of junk files.
You're asked to rate any file before deleting them, specifically on how accurate the files are to their filename/tags rather than rating the content itself. And junk rated files simply aren't enforced on others in the unity cache.

Makes sense, honestly.

This is interesting, Do you people know of more sites like this that are indexed on google?
I love how easy it is to find anime/movies.
>search on google

reboot site:ipfs.io/ipfs OR site:ipfs.io/ipns OR site:drive.google.com

>>51471339
that's interesting

IPFS has a related project called Filecoin, which is their way of incentivising backups/caching

Does this work on windows? :<

>>51471508
It should.

>>51471469
Forgot to take off "reboot"

I'm having trouble trying to conceptualize mutable content like imageboards over this. Could IPNS keep up with active messaging systems?
Given only one host can update the IPNS, it could possibly work out consistently. But it's just really hard to imagine.

>>51469925
>trusted peers
That's just bad practice. They shouldn't be trusted by design.

So seeing as ipfs and gnunet is hot garbage, what will actually replace torrents?

>>51470103
I guess you can serve up chunks of files that are modified but I'm guessing you need a hash collision for that.

>>51470008
>what you're saying is completely pointless encryption
It's entirely necessary for the web to work dipshit. Privacy does not equal anonymity.

>>51471757
why not use both? I see no reason to just drop GNUnet and pretend it doesn't exist.

>>51471757
>>51471779
Neither is meant to replace torrents. They are so much more then that.

nice library. Dark Knight loaded super fast.

What do I need for this to be anonymous? A VPN or more than that

How do you use this thing? How do I download a file from it and how do I find files?

This information should be in the OP.

>>51471973
A VPN has never helped anonymity and has always just been placebo. You need either tor or i2p or another similar service.

>>51472039
>download
Just like if it were a direct download, simply browse to it with your browser. Alternatively, ipfs get <file>.
>find files
either use the index sites posted above or you can e.g. search via google through the ipfs.io gateway (search query site:ipfs.io/ipfs).

>>51469256
Half of the files don't even download anywhere. Pin it pls

>>51471757
as soon as ipfs figures out a way to add stuff without moving/duplicating the data, i'll be adding all my shit to ipfs, so will others i imagine

unlike torrents, everything in IPFS is deduplicated

no more having 100 different torrents that happens to have the same file across several trackers each with their own sets of seeders

with IPFS that file has it's own unique address, which is computed from that file, so it's always the same

one can download that file from anyone who has it, regardless of if the link is for the single file, part of the file, the file in a folder of other files, named something else, appended/prepended to something else, etc. if someone has the /content/ you're looking for, you can use them to get it

>>51472168
also, to contrast with torrents again
i *could* add all my shit to a torrent, but;
1. that torrent would be huge
2. people would need to be aware of this torrent specifically
3. people would need to obtain this torrent specifically (both to get anything and to find out what is even in it)
4. without someone finding/getting/looking through that torrent, i help noone

by adding my shit to ipfs on the other hand, i automatically help anyone who is trying to get data i have, because a particular file will have the same address everywhere

>>51471637
I'm thinking about creating an art dump site, we'll see. I think it works well for "additive" content sites where people upload images which stay there forever, i.e. the content grows but isn't removed. It might be a bit less ideal for websites which change all the time, like 4chan.

>>51472400
Works well in both cases because you don't remove the content - you merely stop linking to it. The issue is actually to get the client to talk to the server when the server should be multiple entities.

>>51472444
yeah, but in the first case people would simply get the "old" site without the content before it gets propagated enough. The other case might be problematic because pages which change all the time might not propagate fast enough and depending on where you are in the network everyone might have different snapshots of the site.

>>51472493
The same issue already exists with http, it's simply not solvable basically. So far it seems that getting the latest version isn't an issue, I can test the new version of my site via ipfs.io and it's there as soon as it loads right after my update.

Is there any way to say... download a specific file (if it is available) by a hash?

>>51472548
ipfs get <hash>

>>51472548
that's literally the point

i know what you meant, but this is the standard addressing method of ipfs, not some sidechannel

>>51467711
>>>most recent talk about it by the dev
>[YouTube] Stanford Seminar - Juan Benet of Protocol Labs (embed) [Embed]
Am I being meme'd on?

>>51472548
I should have classified like an md5 or SHA hash.

What happens if there's a hash collision? I know that's extremely unlikely, but it isn't impossible.

>>51472604
Or maybe it is impossible. I don't know how the hashes are generated.

>>51472548
you can generate a hash from a file without actualling adding/seeding it yourself, then you can use that hash to see if it's available from others

for example, i hashed some file just now which came out as;
QmVgTBEWcbphGeeCjWgnZZD3QrPSNfrz4ctWuB9VTLw6wo
but i didn't add it to ipfs, and since it's a file from a friends' camera, is unlikely to ever be on ipfs
BUT, if i did add that file later on, you will be able to access it from that address, even if i rename it or add it as part of a folder of other files

>>51472598
ipfs uses sha256 currently, it's in a "multihash" format that is open to upgrading later if need be

File: what has science done.jpg (29KB, 512x384px) Image search: [Google]

29KB, 512x384px

What is this and why is it so over-engineered?

>>51472604
It's extremely unlikely, just as BitCoin hash collisions are extremely unlikely. They built in the ability to upgrade hash so, if sha-256 or whatever they are using becomes vulnerable, they can upgrade to sha-4096 or whatever hash is king in the future.

>>51472659
>over-engineered
built for future needs, not just current needs.

>>51472683
So there is no way of completely ruling out collisions? That isn't ideal. It shouldn't make any practical difference, though.

>>51472683
bitcoin uses sha256 as well

if stealing literal money (bitcoin) isn't a good enough reason to throw everything you have into trying to crack it, i don't know what is

>>51471524
Link to windows version?

>>51472704
if you consider what is explained in those talks, you'd realize we do need it now

all that bandwidth wasted man, imagine how blazing fast the internet would be with our current hardware/lines if all static content was on ipfs?

>>51472718
I don't really buy that. If someone cracked sha256, Bitcoin would be worthless almost instantly. You would get a lot of Bitcoin, but no one would want it. You would not make enough money to justify the research.

The much bigger thing is intelligence gathering. The US government uses sha256 for top secret data. That data is way more valuable than Bitcoin and will keep its value even after sha256 is broken.

>>51472704
Well, that's like your opinion, man.

What I see are few ideas squashed into one project. And it seems failed already. It's one year old and doesn't seem really popular.

>>51472777
depending on how you did it, you could get away with a whole bunch of money before anyone found out (you could argue that this may already be happening, mind you)

in any case, the ipfs devs have considered this, which is why they use an addressing format that is not only a hash, but also describes the type of hash used. it's extensible, and can be upgraded later if need be

>>51472546
>The same issue already exists with http
how?

I really don't have an idea about ipfs' performance but I'll have a look into it.

>>51472713
There's never a way of completely ruling out collisions with any hashing algorithm.
https://en.wikipedia.org/wiki/Pigeonhole_principle
There is always a finite hash space and infinite possible combinations of data that can be hashed.

The important thing is to pick an algorithm with a hash space large enough that the likelihood of ever finding a collision is so abysmally small it will never happen.

In the case of SHA256, it would take several trillion times the expected age of the universe of using all the words available computing power to look for a hash collision before one is likely to be found

>>51472812
>What I see are few ideas squashed into one project. And it seems failed already. It's one year old and doesn't seem really popular.

>1990
>first web browser
>year later
>1991
>web's not popular yet, must be dead
>BBS 4ever bro, give it up, it'll never happen

>>51472851
>words
world's

>>51467711
Hi guys,

is there a TDS or TD or somewhere or something I could read?

>>51472812
>alpha software not even remotely ready for production
>already very fast
>already several sites
>already in talks with a major browser vendor
>neocities already runs on it
>ethereum already runs on it
>dead
lol

Will referrer still be mis-spelled for backwards compatability?

>>51472851
Of course. But you could work around it. You could, for example, check if there's a collision and then append extra information if there is.

Anyone know where ipfs stores the result of "ipfs add [file(s)]" ,and the command to remove them?
>>51472887
>Major browser vendor
Sounds exciting, source?

>>51472974
>>51472851
>>51472713

From the creators of IPFS, FAQ on collisions...
https://github.com/ipfs/faq/issues/24

>>51472974
>You could, for example, check if there's a collision
You mean an existing collision?
You want to go and scan through every file that's ever been on the entire ipfs network to make sure none of them have the same hash every single time you add a new file?

>>51472996
>Anyone know where ipfs stores the result of "ipfs add [file(s)]" ,and the command to remove them?
added files are chunked and the blocks placed in ~/.ipfs/blocks
they will be garbage collected automatically unless pinned manually, which simply excludes them from automatic garbage collection
you can remove non-pinned things by running "ipfs repo gc"

>>51473024
>added files are chunked and the blocks placed in ~/.ipfs/blocks
ps. this is just the current type of "repo"
later they plan on a type that can keep file data in-place, so you don't end up with two copies of files

>>51472862
>>51472887
>dead
>failed
>same thing
okay

I2P is not dead, but it failed, too. Tor is more popular, it's not dead, but I wouldn't say it's really successful either.

What I mean, it won't take over the web. Idea is not original, too.
http://www.coralcdn.org/

Does it offer anything so awesome that internet will want to replace HTTP with it? I don't think it can compete with other protocols.

Nonetheless, thank you for pointing out that neocities and ethereum already support it. I wish the project all the best, it's an interesting idea, I was thinking about something similar for some time and I think it will be good to see it get small stable userbase.

>>51473024
Ah, I didn't know gc also removed files you added manually

>>51473061
It can work in conjunction with http, so it doesn't have to be a this or that issue. More of a this and / or that. Which gives it a better chance of gaining a critical userbase. It seems very sound, and has good prospects. I too hope it takes off.

>>51473061
the devs appear to have every intention of making ipfs a big thing

i don't think i2p nor tor really ever were aiming for everyone to be using them

>>51472996
>source
https://www.youtube.com/watch?v=HUVmypx9HGI
OP a shit and didn't copy the link right.

>>51469614
how are you people getting these speeds?

i have muh ports wide open and am connected to 100+ peers, and shit just trickles in.

also, is there any way to change where the blocks are stored? i don't want them on my c: drive

>>51473110
adding a file implicitly pins it as well

>>51473162
Works great on *nix

>>51473162
Might be a config to change where they're stored, not sure.
Download speed depends on how many people are seeding and their speed, as with torrents.

>>51473061
Coralcdn isn't even remotely similar to ipfs in anyway beside the fact that both use p2p tech.

>>51473160
video hangs on buffering.
*wishes it was on ipfs, shit would be smooth. Get with it youtube.

>>51473189
i tried it with a vbox mint install and it was slow as shit as well

>>51467711
How does IPFS compare to Tahoe-LAFS, in terms of hosting a stable pirating network?

>>51473125
Yes, I'm looking at the Alpha Demo on their website and it seems to heavily rely on IP and HTTP protocols.

>>51473142
The problem might be that they will try to please everyone and the project might easily get out of control. Another thing, if the protocol won't take off it might be a big disappointment and core developers could leave. Anyway, these are just hypotheses, so it should be taken with a big grain of salt.

>>51473206
Obviously. If it were the same, it would just be a clone.
Another thing I noticed, the IPFS's IDs seem to be very similar to Tor's hidden services.

>>51473210
I think we have it on ipfs.
Yup, there it is:
ipfs.io/ipfs/QmSd1buim52MtQP
1SH2XhN7FC8c1KdFrF
KBNmgdUKFdss8

>>51473308
Why can't you fags just post links normally?

ipfs.io/ipfs/QmSd1buim52MtQP1SH2XhN7FC8c1KdFrFKBNmgdUKFdss8

File: 1429651709747.gif (776KB, 514x667px) Image search: [Google]

776KB, 514x667px

ipfs add -r /

>>51473299
You can install a DNS TXT record to get a plaintext name fairly easily, which is gneiss.

Also, the ipfs project explicitly tries to be a thin layer that is compatible with other technologies, instead of an all-in-one solution that satisfies everyone out of the box.

>>51473337
Spamfilter triggers 99% of the time.

>>51473338
I don't get it

>>51473337
i mostly get "we think your post is spam" from 4chan when posting them, not always but enough that i just split them up anyway becuase i don't want to deal with captcha again if it fails

>>51473370
add everything on your system to ipfs (he's referring the unix filesystem root with "/")

>>51473392
oh, then I actually got it, it's just not funny enough to be accompanied with a smug anime girl

what is point iof this /?

>>51473415
watch the talk (>>51473337)
it explains the why and the how

are we going to integrate GNUnet into IPFS? the anonymity features of GNUnet might be very useful.

So what's the point? Only point I can see is that you can watch animu as you download it, but with decent internet you download an episode in a minute anyways...

>It's hard to take down
Doesn't matter much, torrents are hard to take down too. But that's not their main approach, no need to take it down if they can just fine you.
>But you can vpn/tor it!
So can you with torrents, and tor is too slow.
>You can have a mutable address (i.e. always points to the latest version of a site)
What's in it for me?
>Peers are found fast for new downloads
Torrents max out my connection

>>51472444
Yeah, the issue isn't hosting content, updating content, or even making sure that content is the latest version. But rather trying to accept updates from users in a decentralized manner.

Suppose you could just route that through HTTP like normal though. Since such posts would have to be given only to the server running software to deal with them anyway.
I mean, it wouldn't be complete IPFS only.
But distributing the running code (To accept posts and update the site) just sounds like a disaster waiting to happen.

>>51473429
ipfs is designed to work with other protocols, it can be run atop tor/i2p

I put TheFalcon's guide on ipfs at http://localhost:8080/ipfs/Qmdwyyu4ifiq5LZ
3gbp4bwREL4Lp9UjepBxdfM
JzRczejq
It's now linked from http://localhost:8080/ipns/QmaGks9KKzu2WykHQjJFJkcUAN4ZoF7ok9h2hXj1WQn47U/

>>51469759
>330 top secret restaurant recipes.zip
wut

>>51473451
ipfs.pics does that already, as an example.
But there are solutions to get a fully decentralized and distributed network, including dynamic content - eris, ecred (I think?), ethereum and more, for instance. An approach like ntpp and using ipfs's direct peer connect is also an option.

>>51473441
see
>>51472168

ipfs uses bittorrent technology, but has huge advantages in addition to that as well

>>51473465
u wot m8? You might have a virus, or your isp might be serving you a redirect.

>>51473452
Is there any documentation on doing this?

>>51473490
It's not "bittorrent technology", it's p2p technology. Huge difference.

>>51473546
But you're wrong. It is p2p, and it was built ontop of BT tech, like breaking up files into chunks and distributing them via hash.

would it be possible to serve html traditionally and pictures/media by ipfs?
That would kinda eradicate several issues.

>>51473595
that's the idea

>>51473606
how do I merge this though? Right now I can only use either/or, not both at the same time, right?

>>51473637
Can't you just use the <img> and <video> tags to embed external media?

>>51473637
i'm not entirely sure what you're asking

>>51473662
>>51473664
so say I serve up a page, do I simply insert IPFS urls to the images for instance? Wouldn't a browser prevent this because of security rules, cross-site bollocks etc.?

>>51473637
Host your HTML traditionally and link your images with a gateway link or the local gateway if you are sure your users are using ipfs
Something like

<img src="http://127.0.0.1:8080/ipfs/QmQJ793JP Har4D8VcTBxE52BztvveuczW 5TMqMt19cV6Ec" >

Hiro should add an /ipfs/ board that would link stuff this way, would be rad.

>>51473637
the game plan includes;
- a javascript implementation, a web dev could include ipfs support into the page itself so it can grab content from ipfs using any current browser
- eventually ipfs support built into the browser as a standard feature, like http is now

i imagine ipfs will initially only be used for static content, and that's fine, static content makes up the majority of bandwidth and ipfs can help a lot with just that

>>51473595
>>51473606
oh shit, that's genius

>>51473715
alright, thanks, that answers my question.

>>51473730
this shit would be ideal for 4chan, just think how many TBs of identical pictures are posted every day.

>>51473745
i guess it needs to be mentioned again that the gateway(s) and there pretty much only for testing purposes, the public ones being so people can try out ipfs with a non-ipfs browser, and local ones for similar reasons

the end game doesn't use gateways, but is aiming for native ipfs support in programs

>>51473590
False. Learn the difference between bittorrent and p2p please.

>>51473790
It doesn't need to be said, doesn't this already work as efficiently anyway? Even though obviously you have to type more.

Alright. Let's build a sadpanda replacement.

>>51473704
See ipfs.pics or neocities for an example of mixing http and ipfs.

>>51473810
a local gateway does give the same result, yes, but wouldn't it be nicer if you browser just supported ipfs itself?

>>51473839
sure, but the problem is that ipfs by design forces users to run a vulnerability nightmare daemon from startup

>>51473808
even the video posted has a graphic with bittorrent + other things = ipfs

You've really sparked my curiosity with this one /g/, but I've got a lot of questions and concerns. It obviously seems like a far better way to distribute the web and in my opinion it seems to agree more with the way the web was intended to be designed in the first place.

First thing I'm curious about is that it seems like, even with widespread pushing and adoption by major players in the tech industry, it seems like it requires a little bit more knowledge to use and navigate for the average internet end user. Is it actually theoretically possible for this to be a drop-in replacement and go unnoticed for the majority of internet users? I'm talking a very impractical and unrealistic update/overhaul of every major operating system and restructuring of web servers; just trying to get a sense of how much impact it would have on a normie's life?

My second question is more about security. You access a website one day, you become a host of that website, right? When that website is updated/changed, do you automatically receive that update and remain a host, or do you have to revisit that website yourself to update your copy and be a relevant host? If the former, how plausible are the existence of exploits that allow for arbitrary code execution on all computers who have visited a potentially malicious website? I'm sure that this was a foreseen potential risk, but I'm wondering how they mitigated against it.

>>51473891
Top kek. You could have least watched the video before opining.

>>51473950
huh?
It was a graphic in the video.
I understand that could be a loose association to help people get it, but if this guy makes that comparison then why can't someone in this thread.

>>51473947
>Is it actually theoretically possible for this to be a drop-in replacement and go unnoticed for the majority of internet users?
Easily: inbuilt browser support + well-populated DNS are all that's really needed. Populating DNSs is entirely server-side and serving content works just as before without change beside making sure to use relative paths.

>>51473947
>Is it actually theoretically possible for this to be a drop-in replacement and go unnoticed for the majority of internet users?
yes, watch the video, the devs are well aware that anything but a drop-in replacement won't happen

>You access a website one day, you become a host of that website, right?
for as long as it's in your cache, yes
>do you automatically receive that update and remain a host, or do you have to revisit that website yourself to update your copy and be a relevant host?
the former, they make it a point that you only get what you explicitly request
this is actually one of the advantages, this behaviour is what makes it "self-archiving"
>how plausible are the existence of exploits that allow for arbitrary code execution on all computers who have visited a potentially malicious website?
how safe is your current web browsers' object cache?

How is IPFS not anonymous?
Tell me how your supposed to trace back a files hash to someones identity.
And even if you do, the identity is also a hash, tell me how your supposed to trace that hash back to someones real identity.

>>51474043
You get the IP of anyone who connects to you. If someone makes a request, they're looking for any person who advertises ownership of the file and directly connect to them. There's no peer bridge as in, say, gnunet.

>>51474043
it's only anonymous in the sense that you can't tell who initially put a file on the network, as everyone is a server

it's not anonymous in the sense that you can tell which ip's have certain blocks by requesting it

>>51474065
>You get the IP of anyone who connects to you.
Literally who cares.
If you can't trace a file back to an IP address, then you can't do shit with it.

>>51474103
They're downloading the file when they connect.

>>51474118
But there's nothing saying the file originated from that peer.

Anyone have big buck bunny link?

>>51474135
There's no other way to have the file than to connect to a peer who has it and to download to it. Thus, if you download from a peer, it's because that node is serving it directly. Serving via cache counts.

>>51474041
Your saying the former but you seem to be suggesting that you meant the latter. Not to belittle your intelligence, but the "former" thing I was asking was that once you visit a website, you keep a copy of that website that gets updated when the website is updated, without you having to explicitly visit the website again and see the updated content yourself. I'm gathering from the rest of what you're saying though that this is not the case.
>how safe is your current web browsers' object cache?
Well my current web browser's object cache doesn't automatically fetch data without me asking for it, which I now gather is not what IPFS does either.

>>51474153
https://ipfs.io/ipfs/QmUW2JxCaELBzy6jhdQjmyvd7sv1EVpunE2kqQta3SekYM

>>51474184
sorry, yea, the latter, you don't get updates unless you visit again yourself
it's 5am here and i'm about to head to sleep

of course the mistake has to be the one that inverts my meaning

>>51474191
Based anon. Seeing this work in VLC is crazy cool.

>>51474168
But he's saying that you can't trace it to the original uploader, you can only trace it to any given person that has accessed it.

If one person uploads child pornography, and 100 people access it, you can only prove that there are 101 people that have the file but you have no idea who the original uploader was.

It's not necessarily anonymous, but it doesn't seem like you can reliably trace anything significant back to someone.

>>51474286
Sure, but that doesn't mean anything to law enforcement.

>>51474286
so they'll just bust 101 people or put you on a list when they're bored.

what's your point again?

>>51473715
>Hiro should add an /ipfs/ board that would link stuff this way, would be rad.
Whoa. That would be a silly boost in content.

>>51474286
I am by no means on the side of LE, but that's definitely not a way they could operate. You can't prove that somebody didn't view the image accidentally. You can prove that the original uploader didn't accidentally obtain and upload it, but you can't figure out who that person is because even someone who clicked on it by accident is a host.

>>51473715
That would be cool as fuck.

>>51474380
Well, this. In case of CP, you just jail everyone. Didn't stop ISPs from sending letters about torrents either.

>>51474416
how long is it on other people's machines though?
Isn't there a garbage collect?
Do I have that demo ocean video on my machine until I delete .ipfs?

>>51474416
Meant to reply to >>51474315

>>51474380
Ok, I get it, being cynical about law enforcement is edgy and cool. But realistically if you think this is a cool idea you have to accept the fact that the government and law enforcement will have a say in whether or not it would be implemented, and they're not going to be satisfied with a model that doesn't foster accountability.

>>51474461
GC happens by default at the 90% cache full mark, which is 9 GB with the default allocation of 10 GB for the cache. Aside from that, pinned content doesn't get deleted.

Rough similarities and differences between bittorrent and ipfs:

-both use content addressing: you ask for a file addressed by its hash, so it no longer matters where you get the data from (and thus you can get pieces from many untrusted peers).
- peers who might have a file are usually found through a DHT overlay network, specifically kademlia. Bittorrent also uses centralized trackers to find peers, but magnet links only use DHT.
- for a single file or block, peers directly contact other peers and attempt to balance downloads across all the peers that already have the file (seeders/leechers)
- re: security, data integrity is guaranteed by content addressing (a bad peer can't serve you a virus instead of the content you wanted) but privacy is not; peers obviously see your IP address when connecting, it is trivial to scrape the list of peers serving or asking for a particular file. The p2p communications are obfuscated (disingenuously called "encrypted" by bittorrent), but the actual data transferred is inferrable from participation in the network at all (open up your own connection to the peer, ask them for a file, and they'll respond). Use i2p/tor to hide IPs.
- Unlike bittorrent, default file block selection in ipfs is sequential, so you can stream files as they are received. Nothing technical in bittorrent prevents this however, see peerflix (popcorn time's backend), or "sequential" mode in qbittorrent.
- ipfs doesn't include file names or file groupings in hash calculation, unlike the torrent metainfo format. This allows files to be shared between torrents. Also, (not default yet), files can have variable size rabin fingerprinted blocks, so data can be more robustly shared between edits of the file. You can hack most of this stuff into bittorrent, but nothing standard has emerged.
- ipfs has a canonical "trust-addressed" mutable data layer as well, ipns, which allows updateable websites. libtorrent has experimental extensions for this, but it's not mainstream.

>>51474503
how is this being edgy and how are people unaccountable? People have been jailed over thumbnails in their fucking browser caches. This is even worse, you're literally redistributing CP when you click on it.

I really don't get what you're even trying to say.

>>51474438
Someone could claim that they clicked on the CP by accident. Even the original uploader would be able to claim that he had merely clicked it by accident and you would have no way to disprove that.

Criminal charges have to be proven "beyond a reasonable doubt" with unanimous decision. Any dumb ass juror or judge can relate to the fact that sometimes you click on things by accident on the internet.
>>51474578
What I'm trying to say is that this gives LE less ability to put actual distributors of child pornography (for example) behind bars because they can't prove that they were the original source of the illicit material. It adds ambiguity to the origins of the file and reduces the legitimacy of the existence of a hosted file to be used as evidence. No one has been jailed just for thumbnails in their browser cache, it was used as supporting evidence on someone who was already a suspect. I think you vastly underestimate how much work prosecutors put into putting people behind bars.

>>51474416
are you a lawyer? that kind of argument wouldn't hold up for shit in court

>it wasn't me your honor! I accidentally clicked on it and my computer redistributed it across the world!

>>51474742
"I accidentally clicked on it" is the part that's hard to believe. "My computer redistributed it across the world" makes perfect sense.

Quick question,

Do you get automatically updates from what you seed?
If that's true it isn't hard to distribute rootkits over the net.
For example i upload cracked software, someone shares it and keeps doing that and i update the bin with a rootkit he other users automatically get that bin and so it spreads if they click on it again lol

>>51473715
that would be cool, but ipfs is too alpha and obscure. You don't want to jump on the bandwagon too early.

>>51473815
You

I like you

Actually, send this shit to sadpandas admin. He would probably be onboard to do it himself (nigga hates DCMA requests)

>>51474547
>both use content addressing
Bittorrent doesn't use content addressing.

>>51474824
They already use P2P anyway. They'd probably be on board.

But sadpanda kind of sucks. The ponies and shitty RPGs and muh sekrit club mentality are all annoying. It has great content, but I'd be on board for a replacement that isn't run by faggots.

>>51474742
I should have appended to my final statement: when the defense has a real lawyer, not some public attorney.

I'm pre-law, and yeah that kind of argument would totally hold up in a criminal court. Your computer literally does redistribute the content across the world after you click on it that's the point of this protocol. As far as the accidental clicking, you don't have to PROVE that you accidentally clicked on it; that's impossible. You just have to ensure that the prosecution can't PROVE that you intentionally clicked on it, which is really hard for them to do in the first place and getting even harder every year the way clickbait media is evolving.

>>51474775
I wondered the same thing. See >>51474041

>>51474880
>ponies
manager is a brony. At least the most cringe worthy thing he does is the title image for ehwiki, while the actual sites are normal

>RPGs
some people apparently enjoy them, and you don't need to use them to use the site.

>sekrit club
was used to keep the feds and IP enforcers at bay. If he makes good use of ipfs he probably won't need it

>>51474547
In summary, ipfs definitely improves on parts of bittorrent, but doesn't do anything that special. Their main contribution is standardizing (or trying to standardize) on unix filesystem-like semantics on top of a content and trust-addressable store. However, the maintainers have intentionally dropped the ball on any sort of privacy improvements, as well as some more efficient p2p algorithms that have developed in between freenet (which is literally ipfs + privacy, but 15 years ago) and now. It's kind of like the language ipfs is implemented in; ignore 40 years of PL research, add in one nice thing (goroutines), and call it good. Not to say that this isn't a smart strategy in general, but it it's kind of a shame.

>>51474836
It does, SHA1. The main deficiency is that file names are included in what is addressed (making swarm sharing less efficient), and multi-file torrents are essentially treated as one giant concatenated file (again making swarm sharing less efficient). The integrity guarantee is still there though.

>>51474975
>was used to keep the feds and IP enforcers at bay
As if it does that. The feds and IP enforcers would have no trouble getting on. The sekrit club just makes it hard to use. It makes it difficult to use any machine except mine and means I can't easily share links with others.

>>51475078
>As if it does that.
feds might be smarter, but it kept publishers at bay until Jewcob decided he needed more shekels

We need a libgen mirror boys, how do we divide the work?

Could you chroot into an environment that you got over IPFS?

>>51473408
I find it is.

>>51475226
I don't see why not. Want to try it?

>>51475329
I would love to if I had a working device beyond my cell phone. You might have to find someone else but I'd love to hear the results if you do.

>>51475226
I'm guessing ipfs fuse mounts with -o noexec because security, but it could work if you override/remount it.

Trying to get IPFS running on ESXi so I can access the node from any device on the network, and leave the node running 24/7. Already have it running on the laptop as a test before I build out a Centos box to host it.

I can get to the web server locally on the laptop through localhost:5001/webui, but other devices aren't able to reach it over ipfs.domain.tld:5001/webui, or via 10.x.y.z:port/webui

Stopped the firewall and disabled selinux, same result. Port scanner sees that port 4001 is open, but not 5001. Added 5001tcp/udp to firewalld and reloaded it, still nothing. I feel like I'm missing something here.

>>51475535
have you tried turning it off and on again?

>>51475535
would you need to change the config so it's accessible on that device? Ie, not 127.0.0.1 or localhost but the 10.x.y.z

>>51474191
>big buck
Stalls at 43MB. Can anyone else confirm?

>>51475853

Probably, but I'm not certain whether ipfs is using the standard httpd.conf/apache2.conf or if it has its own config file somewhere else for however the web server is running

>>51467711
Interesting shit. Watching the vid.

>>51475535
you should have a list of IPs where you can access it from (lan, wan, and local at least), try them directly. Although the interface isn't very useful anyway.

>>51475861
Stopped at 2:14 for me.

>>51476116
Could you $ ipfs get QmUW2JxCaELBzy6jhdQjmyvd7sv1EVpunE2kqQta3SekYM

and see how many MB it can download?

>>51476139

43

>>51476175
Love me some dead swarms

>>51467711

Sounss like a poor mans gnunet

>>51476139
I'm also stuck at 43, looks like it's way underseeded. Gonna download it and upload it so you guys can keep downloading.

>>51476205
GNU means failure in software development anon

>>51476205
More like rich man's given that it's super fast and has content addressing, you can easily assign readable names and it has mutable and immutable addressing, plus an interface that is actually good and doesn't segfault every 5 seconds, and a set of easy to use commands instead of the esoteric gnunet ones.

Really, the only things ipfs is wanting from gnunet is search and anonymity.

>>51474888
Lets make a scenario:
>someone in this thread posts a http://localhost:8080/ipfs/<hash>
>you copy it and put it in the address bar
>you observe that it's cp
>ipfs cached it as you're obviously running a daemon on localhost
You now have 3 options:
1. run "ipfs repo gc" to clear your caches
2. run "ipfs pin <hash>" like a fucking retarded pedophile
3. close the tab and carry on like nothing happened

In case 1:
Probably no one got time to download a single block from you.
Even if they did, it's not likely to be traced to you.

In case 2:
You are seeding CP to others and sooner or later FBI will find you either by someone reporting your IP or by querying ipfs with it's large collection of accumulated child porn. The fact you pinned the file will be used as evidence that you were knowingly committing the crime.

In case 3:
You will get caught in the same way as in case 2.
Now there are 2 more cases:
1. You knew how ipfs works:
In this case, you were knowingly committing a crime and are fucked.

2. You didn't know how ipfs works:
In this case, you might play the card that you didn't know your computer was distributing CP.
However you had to get the CP from somewhere and you'll have hard time proving you just accidentally opened a link on 4chan. They don't have to prove whether you "intentionally" did it because when it comes to CP the possession itself is illegal. The fact you (or your computer) was also distributing it just makes it worse.

>>51475988

The ones presented after running ipfs daemon? nothing there either. Web server is acting like it isn't listening for any external connections on 5001, only local ones.

Issue now is I can't figure out where I need to adjust this

>>51476384
>Even if they did, it's not likely to be traced to you.
NSA asks for that file's seeders constantly the moment they detect a CP file in the network and grab the IPs of everyone that owns it even for one second. Bam, you're fucked.

>>51475861
Yep 43MB stall

File: 12235143_10206894390430970_2827485945480447224_n.jpg (7KB, 240x176px) Image search: [Google]

7KB, 240x176px

>>51476449
Why tho

>>51476139
>>51476175
>>51476196
>>51476116
QmQ747r7eLfsVtBFBSRwfXsPK6tADJpQzJxz4uFdoZb9XJ
New link. That's from the 720p surround version, not sure what the one linked to was.

Hey guys, I quickly made a site that serves as a simple listing of anime ipfs links. Eventually I plan on expanding it to movies and other TV shows. Didn't have much time to put into it today though.

/ipns/QmUqBf56JeGUvuf2SiJNJ
ahAqaVhFSHS6r9gYk5FbS4TAn

I'm lookin for some feedback. What looks good, what doesnt? What can I do to improve the functionality? Feature suggestions? Thanks guys.

Also, all the links are routed through 127.0.0.1, so if you get a pure white background, or the links aren't working, you don't have ipfs running locally. I'm planning on making a gateway version soon. Probably tomorrow.

>>51476418
The NSA doesn't do shit about CP. The FBI handles it.

>>51476552
Looks nice 10/10

>>51476631
Sweet. Thanks man. I appreciate it.

>>51476552
That will work as a simple starting place, but right now it is not at all competitive with existing websites. To come anywhere close to competing with nyaa, it needs to allow user contributions and comments and searching and voting and so on.

>>51476552
You don't need someone to have ipfs running if you use relative links: ../../ipfs (or ipns)/the-hash will always resolve correctly.

Does anyone else get a shitton of traffic even though bitswap doesn't do anything?
ipfs stats bw gives me
RateIn: 22KB/s
RateOut: 14KB/s

but my bandwidth monitor reports 240KB/s down and 140KB/s up

>>51476552
It works, that's nice.
QmYaAHfZUrbS31rZ4wTvENR
TqQMpFsTde29Ldhh6Lkmua2

>>51476738
Also, make sure to use e.g. -t 99999h when publishing so that the name doesn't expire after the default amount of time, 24h.

>>51476738
That's true. I'll start working on that.

Do you know if php will work at all? Considering it's all running on the local host.

>>51476509
Is your daemon up? Trying to get and I don't have any output in my shell.

>>51476552
that's neat

>>51476797
Oh really. That's nice. I'll change the links soon. Thanks.

>>51476803
Thanks m8

>>51476418
>>51476384
I use to torrent CP without a proxy, Nothing ever happend.
I don't think it's as monitored as you say.

>>51476935
Reported

>>51476845
Works for me. I think the ipfs.io gateway is down, try gateway.glop.me instead (which resolves fine) or locally.

>>51476982
Oh wait, ipfs.io also got a hit just now. Guess it's just you then.

>>51476973
Good job, You literally did nothing and broke the rules by saying "reported."
I am not currently breaking any laws or board rules.
I don't even have CP any more.
Loser.

>>51476982
NVM. Randomly started working.

>>51476845
it just started showing progress on my end

>>51477039
>>51477025
I got very slow upload so be patient. Let's bootstrap this shit.

>>51476552
You should try experimenting with the various smart contract frameworks to see if you can get a fully decentralized, distributed website going. It would be a nice proof of concept and a good template for the rest of us.

>>51476552
P gud, pinned.

>>51476982
Holy shit glop.me is great

>>51477406
Pomf clone that runs on ipfs.

Is there a way to "update pins"? For instance, if I want to kill the previous pin and pin the newer version of some content.

My client just stops getting certain files after awhile, but works perfectly fine when I restart the daemon. Anyone else experiencing something like this?

>>51477590
Yeah. the same keeps happening to me on windows :/

>>51477590
Yeah, same issue. I think the client can't handle peers going offline or something like that.

>>51477590
Report it to the devs. Sounds like a bug.

>>51477590
>>51477653
>>51477655
Looks like https://github.com/ipfs/go-ipfs/issues/1896
Might want to comment on the issue.

>>51477475
Im interested in the answer to this too.

If you pin an ipns hash, will it update the pin automatically?

>>51477783
No, it won't. The ipns resolves to an ipfs before pinning, which is why I'm asking.

>>51477788
Hm. Yeah this seems like a problem. How are we supposed to pin the good websites? They are just gonna get garbage collected.

Anyone at all have the answer?

>>51476384
But the point is that this protocol is to be a drop-in replacement of HTTP. The mainstream internet user doesn't even know what IPFS is, is probably gonna be on windows or OSX and not have a clue how to use the command line. This means that the majority of users will pick option 3 because they wouldn't know what else to do.

Imagine a scenario in which someone genuinely did not mean to view the CP and accidentally clicked on it, and is now hosting it. I don't think that this is an unrealistic scenario among the general populace. Are you saying that this person should go to jail? How can you differentiate between who viewed the content with the prior knowledge that it was illicit and who didn't?

The current model is that LE finds the source and they put them in jail. I'm not saying that this is an effective strategy for battling child pornography, but at least it doesn't put innocent people in jail. The only person that you can absolutely KNOW intended to possess the content is the person who shared it initially. With IPFS there's ambiguity in who that was.

>>51476552
Could we get this site in the OP copypasta next thread? I like what hes done there. The potential for it is pretty good.

>>51477863
>differentiate between who viewed the content with the prior knowledge that it was illicit and who didn't?
Pinned v.s. cached

QmQMXctWjsGPi4RhVEF3x
n5XdaF6j8PUBwdH6a4qCdJ6Kj
pls rate

>>51477788
>>51477783
>>51477833
This seems planned
https://github.com/ipfs/go-ipfs/issues/1958

New thread >>51477984

>>51477955
Hohohoho/10