2 days ago i started using peertube for the first time as an uploader since we want to to use it as a sane hoster for a trailer for a show we’re making.
that on one side took some time and energy to get all the on tech/free/libre people in my group to accept that we’re not going to upload to fb or yt and that there still is well be someone watching our video.
so that worked pretty well then and even the thumbnail got fetched into fb. however i never was able to actually play the video on my smartphone ( which is a LOS S3) so kinda old. but apart from that most of my firends even have lower end hardware if at all it doesn’t play on newer ones too. it does partially sometimes.
so after digging in to the problem, and to the whole framasoft thing, seeing peertube is still actually beta, and that your having a crowdfunding campaing, and doing some tests with “my” video.
i really believe that still bandwith is the biggest problem. that peertube as a system has. and since this “people want everything to work immedeatly thing” is very true. i really think this should be priority in any further development. - much more than localasation and subtitles.
maybe there is a way to have even better codecs. but i think there is no way than establishing some systen that is only helping to seed the videos. some server application people could run or this thing with actually backing up data and making it more “loss of server” proof.
(i just spent time on a lot of diffrent peertube instances and i’ve seen the message ( “couldnt fetch , server maybe down”) too often already. in too many diffrent places. and as a user it meant everytime - this thing is not working.
so all in all just a strong pladoyer for making this the top priority.!
also all in all im super impressed with framasoft! i have kinda heard from you in diffrent contextes ( mainly via d*) but i never knew about the full background! you certainly got loads of respects and support wherenver/whenever possible from me !
You make a point, though. Making things work out of the box is important. If you have trouble playing a video, the best thing for us would be that you fill an issue so that we can identify the underlying problem and fix it.
Bandwidth-wise, we cannot really change the codecs (even though better ones exist, of course) as this would reduce greatly the support - especially for phones. It’s really a difficult thing to balance: I dug into the issue when we looked at the open and royalty free codec alternatives available. The support is really uneven, and while we could shim it on browsers that don’t support it natively, it would take a lot more CPU power to decode.
As for the loss of server, I suppose you refer to the replication of a video - to provide redundancy and an additional WebSeed. Maybe we should prioritize it more ; but there are plenty of things yet to be defined before we dig into the implementation. User stories and UX for instance, are going to be tough to define. People already have trouble understanding the federation process ; adding replication concepts on top of that is really gonna mess things if we don’t do it right.
That’s concerning. It would be interesting to see if that’s something due to the software or just because the video/instance is not currently online (in which case, well, we would have to wait for the redundancy feature…).
I think a lot of the bandwidth issue could be solved by automatic mirroring of the data for videos of followed instances. Obviously optional and with filtering.
Sounds a bit problematic to me, as that would mean that hosters need to “waste” disk space, which has become the most expensive thing in the current hosting market.
It also wouldn’t make too much sense, as the problem doesn’t stem from the video itself being too viral or too many videos being consumed, but just the mere fact that (I assume) many PeerTube hosters have bad internet connection.
@thesnake: could you give my fresh host a little try? I’d love to see how it handles some load
