Sometimes your intuition is wrong

Sometimes you get a really cool vision about some gadget in your head and it's really hard to let it go. You have nailed it — it's going to be the next Facebook or Google or Minecraft. Or at least that's what the voice in your head is saying.

You might do even worse if you share the idea with someone. People tend to be supportive and not have the required skill set to evaluate the outcome. Hell, it's most of the times hard to evaluate the outcome even if you're the best person to make the quess.

The only remedy I've ever found to this illusionary problem is prototyping. Sometimes you hit gold, but most often you'll be just proving yourself wrong big time.

This time I'd like to share a failure with you.

The prototype

I had a vision of a game mechanic, where a blind person can take a picture and their "virtual assistant" will call out what's in the picture. I thought this would make perfect sense, and be a great part of the core game mechanic.

Last time I shared a prototype of the "virtual assistant". Today I got the full prototype to work. Finally I pressed the play button and...

Well, test for yourself. WASD to move, space takes a "picture". Try to take pictures of all three animals.

Post-mortem

What went wrong here? Well, as I stated after playing other audio games, there cannot be multiple targets that make a sound at once. It's just confusing as you need to concentrate really hard to distinguish between multiple actors in the scene; "Was that a duck? I seem to be next to a very loud pig, so it's hard to tell".

I do like the way the virtual assistant paints a picture — but I'm also wondering if that verbal painting needs to be dynamic, or if a normal controlled cutscene would do just fine or even better.

Another thing is, that it's really hard to get a full image of what's where using the audible camera. While in a textual representation (check the textual camera demo) you can spot the difference when something changes, but an audio version means that you'll need to poll for the feedback multiple times to get the same information. The bandwidth of audio is just so bad you cannot do it in a push fashion.

All in all it was a good iteration, and I'm glad I didn't spend any more time on it — two days wasted could have easily been two years and a stack of money. Now I'm getting back to the drawing board with the next ideas for the story.