If I understand well Jitsi is written mostly in Java which sounds scary.
Would it be improved if it was written in another better language ? We all know how much resources and problems this language have everywhere.
For a matter of comparison I used to use a streaming software written in Java and despite the fact it worked as expected it consumed A LOT of resources from the host. As soon as I could replace it with another software written in C/C++ the extra performance was unbelievably better, and simpler to manage
Can it be considered to be re-written in a proper language in order to perhaps improve performance on how many simultaneous people the same hardware can handle for both cases ?
Which c++ software did you use?
Why don’t you make a fork with C++ baseline on github?
If I am writing to talk about this issue how do you expect me to do such a thing ? If I was able to do such a think I wouldn’t write the above, I would just fork it in C++ and post that it is available.
I am writing it to the people who write Jitsi, if you didn’t get it.
You criticize Java but you’re not skilled to criticize it?
Wait a sec… so you’re not able to code it in C++? I thought everyone was able to do it… hmmm maybe that’s why some people still code in Java
Anyway which was the C++ software?
Every time developers feel criticized they come with the same stand saying “Write the code for me” as their main defence for not having anything else to say or add.
Don’t need to go much in deep on how bad Java is for everything in general. World knows each day how Java brings more problems than solutions.
And differently from what many defensive devs say people who put up the problem don’t need to have the will to code that, just to understand what they are talking about. Even if I could I don’t need to do.
Need to concentrate in discuss the matter and the potential solutions for it rather than try to take the issue out of the sight.
Sure you can have it. At this feature level rewriting it to a different language would take months for multiple devs, so if you have 2.5 mio $ laying around you can hire a team to do that for you.
Sincerely I hate Java in the deep of my soul…
Anyway you’re still avoiding my question, which is the software made in C++?
Just consider few things… Jitsi one day will use mainly WebRTC with no server side video bridges, becasue will become useless since the technology will be evolved and spreaded enough to let people make 3+ meetings with no need of a centralized server over for handshaking, not right now unfortunately.
Recode and entire software from one language to an another it’s a huge and long work… plus if you consider that you have a dozen different way to create a web socket in C++ and anyone use a different one, plus keeps hundreds lines of code… you’d probably understand why they used Java.
remember to say which is the C++ appliance!
Then share it as OpenSource and wait for people asking for help and criticize why you didn’t made it in Assembly since it’s more performing
Sorry I didn’t avoid it, just missed it.
The streaming software (it isn’t videoconferencing) was simply nginx with some extra stuff but it also played a key role in doing a lot of caching in front of it at well (which replaced the heavy frontend services written in Java).
The work CPU usage was reduced in a proportion to 1 to 6. So the work before done by 6 servers was then done by a single one.
Regarding recoding you are right about being a huge work, but perhaps it can be done in modules, by parts, starting on the ones that consume more CPU.
Good you hate Java.