Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jul 06 11:46
    devxpy commented #69
  • Jul 06 01:18
    RangerMauve commented #69
  • Jul 05 22:08
    devxpy commented #69
  • Jun 24 20:57
    joehand commented #88
  • Jun 24 20:56
    joehand commented #88
  • Jun 24 20:55
    joehand commented #88
  • Jun 24 20:55
    joehand commented #88
  • Jun 24 20:54
    joehand pinned #88
  • Jun 24 20:54
    joehand closed #88
  • Jun 24 20:54
    joehand commented #88
  • Jun 24 20:47
    RangerMauve commented #88
  • Jun 24 20:39
    todrobbins commented #88
  • May 31 16:11
    RangerMauve commented #88
  • May 31 08:14
    decentral1se commented #88
  • May 30 21:59
    todrobbins commented #88
  • May 30 21:56
    todrobbins commented #88
  • May 30 21:41
    RangerMauve commented #88
  • May 30 21:28
    todrobbins commented #88
  • May 30 21:26
    todrobbins commented #88
  • May 30 21:22
    RangerMauve commented #88
dat-bot
@dat-bot
void09 does it have to have an ip/domain, or can it be dht?
void09 dht-like*
ctOS void09: I wrote an article yesterday that you should read. https://www.ctrl.blog/entry/dweb-readiness-runet
ctOS Looks at the centralized points and peer discovery in Dat and IPFS.
void09 yeah i actually read part of that :)
ctOS Read the rest. Look at the previously published article for info about uncensorable domains.
void09 The Dat project also has a single-vendor dependency for DNS resolution as the project defaults to using Google DNS over HTTPS by default. - why ? :
dat-bot
@dat-bot
void09 IPFS wouldn’t remember the previously connected DHT nodes and would be just as lost. - interesting
ctOS void09: they want to use DNS over TLS instead of the regular unencrypted decentralized DNS architecture.
Aral Balkan
@aral
Hey @ct0S, thanks for the article. Btw, I noticed this message on your site: “Ads on this site don’t track or stalk you. Please disable your blocker.” So I looked to see what ads they are (as our Better Blocker was blocking them and we don’t block non-tracking ads) and saw that they were Google. Am I missing something?
@jhand On the Dat vs IPFS page, it would be good to note that Dat is supported by a not-for-profit whereas IPFS is supported by venture capitalists (and they will need to make a few billion dollars at some point… how exactly?) :)
Aral Balkan
@aral
@RangerMauve Indeed, with p2p you can build a basis for an ethical economy but it will not have the same success criteria as a centralised system. For anyone interested in building a p2p system for purposes of furthering social justice, making a billion dollars is a failure state :)
dat-bot
@dat-bot
void09 ok, so dat does not do content based addressing for files/chunks of data?
dat-bot
@dat-bot
void09 IPFS and Dat share a number of underlying similarities but address different problems. Both deduplicate content-addressed pieces of data and have a mechanism for searching for peers who have a specific piece of data. Both have implementations which work in modern Web browsers, as well as command line tools.
RangerMauve
@RangerMauve
void09: Dat doesn't do deduplication yet. 😅 It's possible with how the protocol works, but at the moment I'd you add a file to an archive multiple times, it'll show up multiple times in the content log.
dat-bot
@dat-bot
void09 ineteresting. so it seems neither dat nor ipfs is perfectly suitable for my purposes
void09 ipfs would need an extra application layer to do versioning and maybe .diff sync between peers, or just let it redownload the full file maybe.
void09 but i still don't get one thing
void09 the same files between two dat repositories/sites, cannot be cross-shared, right?
void09 the same file*
RangerMauve
@RangerMauve
void09: Nope! Each dat archive uses a separate swarm of peers. This gives you higher privacy guarantees. You don't know what data a person has unless they share the Dat URL with you.
dat-bot
@dat-bot
void09 how would you know that in the case of ipfs ? crawl the whole dht and make a database ?
void09 you can't directlY enumerate an ipfs' peer pieces/files, right?
Aral Balkan
@aral
Thinking out loud: to avoid brittleness in multiwriter hyperdb, I’m thinking of keeping the purpose/contents of the origin hyperdb/hypercore as simple as possible: create it and then immediately create a separate hyperdb and authorise it and that’s the only thing I will ever use the origin hyperdb for. (Use case: reducing potential for corruption when having to reproduce the origin node from a read-only backup, etc.) Thoughts?
  • origin = original
RangerMauve
@RangerMauve
void09: crawling the DHT is what people do for BitTorrent. IPFS has a similar story, and I think the block service lets you get data pretty easily
Aral Balkan
@aral
Also just reopened mafintosh/hyperdb#163 as we do need reproducible writers on hyperdb if we want to be able to recreate writers after they’ve been destroyed (e.g., after a “sign out”) so that people do not have to re-authorise a previously-authorised node.
@void09 Where does Dat use Google DNS over HTTPS? (And thanks)
dat-bot
@dat-bot
void09 @aral that was from the article ctOS wrote https://www.ctrl.blog/entry/dweb-readiness-runet
ctOS aral: non-personalized ads without tracking by Google. They do use cookies for limited purposes like rate limiting, and fraud prevention (the same purposes allowed by the GDPR)
dat-bot
@dat-bot
ctOS aral: it's the type of ad network designed to work without user consent and still comply with the GDPR.
dat-bot
@dat-bot
ctOS void09: to be fair, only Dat really works on the web and only in one browser. IPFS jams everything from every source into one origin. It strips away almost all security layers of modern browser. ZeroNet has the same problems yet they encourages users to do PGP encryption within one shared origin. It's madness.
dat-bot
@dat-bot
void09 well it is called inter-planetary filesystem
void09 it's a feature, not a bug :P
void09 designed to eliminate useless fragmentation of sources for data
void09 what I thinik it lacks, and pretty much any p2p programs lacks this (correct me if i'm wrong), is some sort of peer sorting by a metric such as routing distance, or geographical proximity
void09 especially useful in large swarms
void09 I find myself in Europe, downloading from hosts in brazil, on a torrent with hundreds of seeds :\
ctOS It’s quite useless for an application resembling the web with individual websites maintaining their own security contexts.
void09 ok so the danger here being, anyone can see that your ip used/hosts a certain reouces on ipfs, right ?
void09 privacy leak
ctOS void09: no. a webapp, say a chat app running on IPFS, share the same origin (localhost:8080) as any random other web page you read in your browser.
ctOS So any random webpage can read your chats.
ctOS Security contexts on the web are [to varying degrees] isolated by origin/domain.
void09 I never thought of that.. is it really possible with the current ipfs?
dat-bot
@dat-bot
ctOS Of course. It’s not magic.
void09 ok well, what can I say then. don't use it for chats :P