Restructuring TouchCORE for CLI

[Calem Bendell]

I was thinking of making an alternative UI for TouchCORE over the course of the summer.

In order to hook it up to TouchCORE, there would need to be a CLI version of the program. Is there still any interest in doing this? It was talked about a few times, but I'm not sure how serious people were about it. The upside of restructuring the app a bit is that it would make debugging and GUI programming a little easier, but it would take maybe more time than it is worth, and the project is a research project at its core.

I was going to make a TouchCORECLI branch May 11. If there is an interest in doing this and I am able to do it, will people be available over the summer to help?

[Jörg Kienzle]

Hi Calem,

I don’t really understand what you mean by AI for TouchCORE. And what is CLI?

• J

[Calem Bendell]

Oh boy typo, just fixed it!!! I meant to say UI! ... not that an artificial intelligence for TouchCORE wouldn't be cool.

CLI for command line interface to the program, making a way to access functionality that doesn't depend on multitouch 4 java.

[Matthias Schoettle]

The model-related part does not depend on UI. So there is an API that can be used. What kind of UI are you thinking about? Have a look at the "Architecture Overview" page in the Wiki :)

[Calem Bendell]

My expertise for UI stuff right now would be Unreal Engine and a little HTML5, so I naturally lean toward these.

Both of these support some amount of cross platform multitouch. Unreal Engine 4 was just released free for noncommercial projects and is open source, and it is really mature and would be much more performant, but it might also be harder to program than HTML5. An advantage to working with Unreal Engine is that GUI components can have a sort of physics, making the self organisation of charts the result of a step by step in a physical simulation. The GUI can also be made with a visual node-based programming langauge for the most part if using UE4, which might be easier. Using either UE4 or HTML5, I was hoping to take advantage of the expanded capabilities of these rendering engines to experiment with better ways of visualising and organising models rather than static 2D renders.

[Matthias Schoettle]

We are envisioning an HTML5 interface with a server-side backend that utilizes the existing infrastructure using EMF. I think development time of new functionality would be faster on the UI side and allows more advanced features. This would also support far more devices/platforms than right now.

Also, this hopefully would allow easier integration of some model repository that everyone can access (ala models in the "cloud").