Practical multiplayer framework (verse protocol)
Posted: 19 Nov 2013, 07:58
It uses a C/C++ API and a tree data model. I don't know how your data model works now, but this is probably the easiest way to sync it across different clients. It's got pretty good compression now days. It was originally used in the psudo-mmo "love". Back then it didn't have very optimized performance compression, and it still managed pretty well.Verse 2.0 is network protocol for real-time sharing of 3D data. It is intended mostly for graphical applications of colaborative virtual reality. It could be used for sharing data between applications like Blender.
The maintainer (jirihnidek) has been very responsive to any issues that I had with the python API.
It seems like a pretty good fit, although the documentation is a bit crappy. I don't know how much it would take to integrate into opemmw. Probably not a high priority, but it would allow for some interesting functionality down the road.
I've been trying to promote the verse protocol lately, because with VR picking up we need an underlying protocol, and I don't want it to be unity.
Brainstorming things that could happen post 1.0
An advantage to verse is that you could write anything (say web apps) and have them interact with your verse scene. It's just a hop-skip-jump from verse co-op to dynamic MMO with auction houses and dynamic content. You could write python or javascript that interacts with openmw scenes using the verse api. So a python dungeon generator or a javascript auction house.
It would also make it pretty simple to integrate opencs with a running instance of openmw, which would be cool.