Hacker News Reader

Comment by denisnazarov | Hacker News Reader

This is not actually p2p. Check out https://github.com/feross/webtorrent. Hoping to hear feross' input.

Does that matter? The joker.org server is doing the p2p download bit and streaming the content to the browser.

It's likely that it's caching material for torrents. I tried Guardians of the Galaxy and could skip almost all the way to the end of the film when it had only just begun.

That could quite possibly just being taking advantage of the torrent block downloading. Normally you download blocks based on the least available, but you can pull them in any order you want. This would allow the server to selectively buffer whatever you need fairly quickly.

That said, its likely that they're cacheing.

Cache is likely and that has legal ramifications. If they are just providing a viewer their hands are clean legally, if they cache they are in danger RE DMCA. Cool product, works remarkably well.

I'd suspect they have to be using a cache, too, but I tried a whole slew of random torrents off kickass.to, and if it played it at all, I could scrub every single one of them to any point in the video with a mere 1-2 sec delay. So, they've either cached a gi-normous amount of torrents, or they've come up with something really interesting in the ability to dynamically stream any block with very low latency.

Good testing scheme

A free seedbox with no login barrier is unscalable for many reasons. Its neat, but I would rather see everyone who thinks this is interesting contribute to the Webtorrent project, which strives to make being a torrent node as easy as visiting a URL, using WebRTC as the data channel.

>I would rather see everyone who thinks this is interesting contribute to the Webtorrent project

Um. I'm not quite sure the set "thinks this is interesting" is entirely included within "has the ability to contribute to the Webtorrent project". In fact, I think the former more than likely has a bigger intersect with "will probably break the Webtorrent project and/or disrupt the community" than with the positive contributor set.

webtorrent doesn't seem to be getting anywhere. The most important part, client integration, hasn't occurred anywhere, and the author hasn't even pursued it.

How can you say it's not getting anywhere? Last time I checked, Feross is pushing an incredible amount of modular, high quality code. His #webtorrent IRC room is highly active and has even become a meeting place of sorts for people working on other decentralization projects. The idea that client integration is the most important part is entirely your opinion, and it's an opinion that not many agree with. In a completely browser-based form webtorrent could still have the potential to cut media hosting costs by a factor of 100.

I explained why in my comment. The protocol needs to be integrated into desktop clients and it hasn't been. Moreover, feross is not trying to convince client developers to incorporate webtorrent, nor is he submitting patches to those clients.

foobarqux, actually he is planning on doing that. I met with him just the other week, so I know this for fact. It just takes time, maybe you could help contribute?

Reply on news.ycombinator.com