Well this is fun : ). I'm "dudeInBlue" in the chat.
dudeinblue, you should have talked more! we were haroldandkumar here at the tokbox office
I (dudeinwhite) attempted to talk but I kept bugging out :(
Show HN: Multiparty WebRTC video chat
41–50 of 50 posts
Re: Show HN: Multiparty WebRTC video chat
#42Isn't the point of WebRTC to avoid dependence upon a company like OpenTok?
Re: Show HN: Multiparty WebRTC video chat
#43Isn't the point of WebRTC to avoid dependence upon a company like OpenTok?
The point of WebRTC is to allow real time communication in the browser. You can do this via p2p data streaming without any services but you are limited to p2p and its shortcomings (like maximum of 2 users, blocked ports, firewall, etc).
By using a service like OpenTok, you can augment your WebRTC experience because you well be able to stream to thousands of users, tunnel through blocked ports, and generally more firewall friendly.
Great question tho.
Re: Show HN: Multiparty WebRTC video chat
#44Isn't the point of WebRTC to avoid dependence upon a company like OpenTok?
The point of WebRTC is to allow real time communication in the browser. You can do this via p2p data streaming without any services but you are limited to p2p and its shortcomings (like maximum of 2 users, blocked ports, firewall, etc). By using a service like OpenTok, you can augment your WebRTC experience because you well be able to stream to thousands of users, tunnel through blocked ports, and generally more fire…
I didn't know that webrtc was limited to two users. If it is how does opentok get around that? My understanding of webrtc was that it uses UDP so it can get around firewalls and such by using STUN or things like that. So my experiments with webrgc used an open source library that was using some Google sever and the author of the library's server for making connections and also firebase. I think k the guys name was muaz Kahn or something. Anyway I think all you really need servers for is the initial connection establishment. And there are free servers up providing that. So if webrtc truly has a limitation on 2 peers then I doubt opentok actually has any magic patch for webrtc.. maybe they have a script that creates a bunch of Iframes or something.
Re: Show HN: Multiparty WebRTC video chat
#45We use opentok a lot and are developing a system for it. My biggest issue with it is reliability and incorrect behavior on the opaque opentok side of the libraries. iOS support on the iphones seems unreliable right now and has a bunch of unspecified silent errors when you init a subscriber from a new stream. Even with the javascript libraries we sometimes find ourselves in situations where the best solution is to programmatically refresh the page because the opentok errors become unrecoverable. The javascript side is better though. However, if tokbox.com is blocked by the firewall right now it does not seem to attempt to route around it.
Right now I think opentok is still one of the best webrtc offerings in the market but that is mainly because of the mobile support race which they seem to be losing ground on.
My one big request is to make the ios sdk less opaque. Just give us the source of the ios sdk. If it is going to be broken and have a bunch of issues that sucks but as long as I can start digging around into the sdk and fix my issues I care much less. As it stands right now it does not support iphone 4 and below and currently it is not showing the subscriber video on an iphone 5 even though the code works perfectly on ipads. Since the sdk is a blackbox and is the source of the silent errors all I can do is throw a bunch of weird hacky try/catch solutions around it.
TL;DR - Release the sourcecode for the ios sdk because it is buggy and the developers who use it need to route around the bugs.
Re: Show HN: Multiparty WebRTC video chat
#46Nice job, guys! This could seriously put skype and webex (ughhh) out of business.
meanwhile google hangouts still requires a plugin. Even in Chrome.
Re: Show HN: Multiparty WebRTC video chat
#47Re: Show HN: Multiparty WebRTC video chat
#48Link to source code: https://github.com/songz/OpenTokRTC
Correction: Link to API code to connect to OpenTOK, not source code as you would expect...
Re: Show HN: Multiparty WebRTC video chat
#49Earlier quoted context omitted.
The point of WebRTC is to allow real time communication in the browser. You can do this via p2p data streaming without any services but you are limited to p2p and its shortcomings (like maximum of 2 users, blocked ports, firewall, etc). By using a service like OpenTok, you can augment your WebRTC experience because you well be able to stream to thousands of users, tunnel through blocked ports, and generally more fire…
I didn't know that webrtc was limited to two users. If it is how does opentok get around that? My understanding of webrtc was that it uses UDP so it can get around firewalls and such by using STUN or things like that. So my experiments with webrgc used an open source library that was using some Google sever and the author of the library's server for making connections and also firebase. I think k the guys name was mu…
To use webrtc with more than 2 users you would need a webrtc streaming server that receives an incoming webrtc packet and sends it to all listening connections. Like a pub-sub model
Re: Show HN: Multiparty WebRTC video chat
#50We use opentok a lot and are developing a system for it. My biggest issue with it is reliability and incorrect behavior on the opaque opentok side of the libraries. iOS support on the iphones seems unreliable right now and has a bunch of unspecified silent errors when you init a subscriber from a new stream. Even with the javascript libraries we sometimes find ourselves in situations where the best solution is to pro…
Dev Evangelist from TokBox here. I'm sorry you had difficulties working with OpenTok. Our technology is bleeding edge and is moving terrifically fast so a few bugs are a tradeoff thats worth the ability to be able to innovate in the market. What this means for us as a business is that we've doubled down on our support talent and infrastructure, all of which maintain quite a high level of responsiveness to issues that come in (our premium support customers seem just as happy to have us as we are to have them). I'm curious if you've used any of those channels and if your issues went unanswered - if there's a leak in those pipes that would be very important to us to fix. Lets address those silent errors when initializing subscribers, please email support@tokbox.com.
Open sourcing the technology isn't really an option for us, and its not just because the technology we are working on is our IP. Laying the infrastructure and process conventions for accepting patches outside the company while driving a growing internal engineering team is non-trivial. Making a repo public isn't some magical pill that will help you get to your answer faster, there's a lot of planning and documentation needed to make open source work. If anything I'm confident that the number of inbound support queries would increase because of questions about the code itself. This would slow down the response time for developers that truly have an issue we should address quickly. An open source project could work under these conditions but for a team of our limited size it would only slow things down.
That being said we love the community of developers that contribute to open source and we have been trying to do more of it ourselves. Our framework integrations are OSS (PhoneGap plugin, Titanium module, more to come), and we're eternally grateful to participate in projects like Google's webrtc implementation.
TL;DR: Open sourcing the project would slow the development down. If you need support, we're all ears and want to make that experience the best it can be (and the forums are public so you can hold us accountable).