If you want to generate interesting questions about decentralization, just stare at how modern BitTorrent works and try to explain why the tech/ecosystem are the way they are. I'm still struggling to construct a full narrative
-
-
The decentralized search experience was never particularly good. Tribler took forever to load, there weren't many total results, and results had little information- and weren't particularly high quality (a lot of broken results)
-
I think I agree, but I haven't used them all. For example, Perfect Dark's "DKT" + DHT combo https://en.wikipedia.org/wiki/Perfect_Dark_(P2P)#DKT+DHT+DU …. Also, techniques which scrape the DHT locally then create a fast local index. OpenBay also may fit I think https://github.com/isohuntto/openbay …. BTDigg also a partial fit
-
I think the claim that decentralized search was just a worse experience is a decent default assumption if for no other reason than creating a good decentralized UX seems to be an order of magnitude harder than an equivalent centralized system.
-
Distributed systems throw out a lot of the assumptions that make for good user experiences- e.g. you don't have a bounded amount of time before seeing search results, so things have to trickle in out of order.
-
Yup. I also think it that completing 80% (as in 80/20 rule) of a product for a centralized app is pretty usable on avg, maybe a bit buggy. 80% for a decentralized app is mainly spent on getting it to work securely and the UI doesn't get enough attention (Tribler may be a good ex)
End of conversation
New conversation -
-
-
ThePirateBay & other centralized search engines also we're fundamentally a very familiar paradigm. They ran in the browser, like Google. We don't really have good example of search outside of web browsers.
-
True. We could still have in-browser decentralized(-ish) search. Another user mentioned BTDigg in this thread. Also Magnetico (https://github.com/boramalper/magnetico …). The previous generation of file sharing did have in-app search, but we just know that is a legal nightmare. Look at PopcornTimepic.twitter.com/nMwuhSC3Ng
-
I can't really think of any self hosted app that anyone uses (the best example I can come up with is Jupyter). There isn't very much decentralized stuff that's been done with WebRTC (since most use it with a special signalling server)
-
Jupyter is a good example. Electron apps can probably be viewed as in a similar camp just not literally opened up in the browser. You could also build this stuff as standalone frontend apps hosted on IPFS that use local storage aggressively
-
It remains to be seen whether a truly controversial app hosted on ipfs won't be successfully taken down from public gateways.
-
Yeah, but I think it seems promisinghttps://twitter.com/backus/status/1012545688633106432 …
End of conversation
New conversation -
-
-
Aren't centralized trackers easier to monetize? Even if $$ isn't primary motivation, as litigation increases, only sites with funding can defend and stay up? Just an idea
-
Yeah, but monetization didn't always hold back p2p file sharing tech. Monetization also doesn't seem to motivate a lot of other centralized torrent websites on the internet. World of piracy has hard to explain incentives IMOhttps://twitter.com/backus/status/1006391268216205312 …
End of conversation
New conversation -
-
-
DHT/PEX made the process for finding peers slow and unreliable especially for the long tail of things people torrent. The fact that it was structurally opposed to private trackers, where most of the more devoted users flocked to certainly didn't help.
-
Agree it is slower. I think it mainly exists as a demotivating force for law enforcement, reminding them that taking down torrent trackers will only slightly degrade the experience, not stop anyone. Your point about private trackers is very very interesting!
End of conversation
New conversation -
-
-
Very good question. It seems like torrent tech stalled somewhat. We are still using the same UI from a decade ago. Pirate bay doesnt even catch spelling mistakes. No thumbnails or torrent metadata like IMBD or trailers.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Search and trackers are still centralized because that's a lot harder to build in a p2p way (without being spammed, DDoSed, etc.), although this will change once we see protocols with built-in cryptocurrency micropayments.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.