How to Build a holographic app that will also run on Windows desktop

holodoctor

Hi,

I am thinking of starting a holographic (3D) app project using C#/Unity. Since there is very few HoloLens devices out there, I am considering an option of creating a universal app that will run both on HoloLens and on other Windows devices at this point. (Same binary, or same codebase with minimal customization for each platform.)

(1) So, first of all, is it even possible?

(2) If so, what's the best practice?

Just to be clear, I can see that some types of holographic apps will not make sense outside HoloLens. But there are also types of apps that can be optimal on HoloLens and yet still can be usable on desktop or phone as well. (Many of the games, I imagine, belong to this category.)

Has anybody tried developing holographic apps for universal platforms?

Thanks,
~h

Patrick

I had a misconception on how Unity cobbles these files together. The UNITY_HOLOGRAPHIC define only applies to the main project, so you can use it in app.cs, but not in your scripts because Unity reuses the projects that contain your scripts for both solutions. I think Jerome has a better answer to this.

Jerome

I would recommend to create a specific unity project just for Hololens.
has the project must be compiled in x86, limited to 2Gb of RAM and has to support a specific camera and other gaze / voice / gesture behaviors, its better to have a dedicated project.

Patrick

1) yes

2) I don't think we have published best practices. I've honestly not tried to do this (except for the last few minutes ).

All that being said, I'm sure you've accidentally run on local machine instead of remote machine or emulator, and you probably noticed that the app works. You may not be able to move the camera on the local machine, but if you have a script that controls the camera with WASD, the camera will move.

If I were going to try this, I would build and export two versions of my app from unity. One with enabled for VR/HoloLens checked and one that isn't. I know, not the best, but...the rendering and input profiles are going to be quite different anyway. If you do this then you can put HoloLens only code in #ifdef UNITY_HOLOGRAPHIC, and you can put the non-HoloLens code outside of these #ifdefs.

holodoctor

Thanks, @Patrick. Let me try out what you suggest. ~h

Jerome

check this:
http://stackoverflow.com/questions/30479171/windows-10-uap-determine-if-device-is-iot-e-g-raspberry-pi-2
and this
https://forums.hololens.com/discussion/1048/analyticsinfo-deviceform-vs-apiinformation-istypepresent

you should test at runtime if the device supports some functionalities and change the behavior of your app based on the result.

holodoctor

@Patrick, I tried your suggestion, but I couldn't make it work. For example, I tested this simple script attached to a cube.

public class RotateCube : MonoBehaviour
{
    Transform cube;

    void Start () {
        cube = GetComponent<Transform>();
    }

    void Update () {
#if UNITY_HOLOGRAPHIC
        cube.Rotate(0.1f, 0.2f, 0.3f);
#else
        // cube.Rotate(0.1f, 0.2f, 0.3f);
#endif
    }

}

Unfortunately, the flag UNITY_HOLOGRAPHIC is never set in the generated VS project (under UWP) which includes this script regardless of the VR/Holographic player settings.

Thanks,
~h

holodoctor

Thanks, @Jerome. I do have some experience building UWP apps, but I'm rather new to Unity-based app development.

There are many issues to deal with when you develop an app for multiple platforms, but in terms of code branching/conditional compilation/build, there seem to be three options. To summarize what I have explored so far,

(1) You can use the standard UWP techniques, as @Jerome suggests. For this, I think you will need to use NETFX_CORE or WINDOWS_UWP in Unity scripts. I couldn't figure out what are the differences between these two constants.
(1A) You can check the value of VersionInfo.DeviceFamily. Apparently, this use is discouraged by Microsoft.
(1B) The recommended way is using ApiInformation.IsTypePresent(). I am not sure how useful this will be, at least at this point without much Unity dev experience, since I will end up mostly programming in Unity API and I will not know generally how they are mapped to UWP API.

(2) Per @Patrick's suggestion, I tried generating two VS apps (from the same Unity project) by changing the VR-enabled settings. This actually "works". Sort of. It enables/disabled the flag UNITY_HOLOGRAPHIC in the VS app projects, but not in the lib project where Unity asset scripts are included (Assembly-CSharp under the UWP folder). So, again, this option is not much useful to me.

So far, I haven't found any optimal solution. Please let me know if anybody has a comment or a different suggestion.

Thanks,
~h

Patrick

I had a misconception on how Unity cobbles these files together. The UNITY_HOLOGRAPHIC define only applies to the main project, so you can use it in app.cs, but not in your scripts because Unity reuses the projects that contain your scripts for both solutions. I think Jerome has a better answer to this.

Jerome

I would recommend to create a specific unity project just for Hololens.
has the project must be compiled in x86, limited to 2Gb of RAM and has to support a specific camera and other gaze / voice / gesture behaviors, its better to have a dedicated project.

holodoctor

Thanks for the input, @Patrick @Jerome.