[jamii]

I did a fair bit of work in Unity earlier this year. The Unity UI makes a lot more sense once you realise it is built by the objects in the scene. You can code entirely in Visual Studio and just use the Unity window to view the results. But you can also start extending your entities with custom controls and debugging panes and build your own graphical tools. Unity basically makes it incredibly easy to whip up graphical debugging and editing tools for the times when that makes sense but doesn't prevent you from working in whatever IDE you are comfortable with.

For example, I worked on a tool for procedurally deforming enemy meshes to prevent the whole 'attack of the clones' vibe that many games fall into. Rather than use print statements to figure out what was happening in the code, I used Gizmos [1] to draw various important coordinates on whichever mesh section was under the mouse and added handles to let me interactively modify the inputs to the algorithm. I also added debugging output on the side panel that updated in real time as I deformed meshes. In Unity its trivial to build those kinds of custom tools. Debugging the same problems with the usual print statements and compile/restart cycle would have been much slower.

I take issue with a lot of the other design decisions in Unity, but the UI is a massive improvement over most work-flows once you learn how to customise it.

[1] http://docs.unity3d.com/Documentation/ScriptReference/Gizmos...