Hi, I am a member of Membrane Core Team and I have created a GitHub discussion to aggregate ideas what could be done membrane_core v2.0 someday.
If you are using Membrane and you have some ideas, about what we could change in the framework (I am not talking about specific plugins ), feel free to write your thoughts down.
membrane_core v2.0 will bring API breaking changes, so any ideas on how to modify the framework API to do something will be appreciated.
If something just bothers you in the way Membrane Framework works, you also might write it there.
You don’t have to be a member of our team to share your thoughts, we especially appreciate the feedback from the community.
If you don’t have any specific ideas, but liked some of the propositions from the discussion, feel free to upvote them, so we will know what people from the community think.
Not an idea about Membrane itself, but one thing I would suggest to any team putting out a framework or OS project is to make it easy for people to adopt it (or the latest version of it) by means of screencasts, guides and tutorials etc
Yeah, along those lines, making sure there is an example in the demos repo for every major use of the various element plugins. That would be nice.
For 2.0 stuff?
I wonder if Membrane should use GenStage? Or are there problems with that? An Element seems like a more tightly specified GenStage.
There isn’t actually very much that is “media” specific about elements. I have done a pipeline from audio to text to other text to audio again.
I would actually love a more thoroughly specified variant of GenStage that is introspectable at runtime. It can be a general pipeline tool for Elixir whether you run it with Membrane, Broadway or Flow.
Providing the schema for buffers could be a nice option in service of that. More self-documenting if they aren’t just binary data.
But most of this is not in direct service of Membrane’s goals so I am not sure it should be a priority.
), feel free to write your thoughts down.
