About recursive frog (twitter: @RecursiveFrog)

My Ludum Dare 48-Hour Compos
LD#22 -- HUGMONSTER (-o^_^)-o (Unity Web Player)
LD#23 -- Burning Platform (Android app (written on an Android tablet))

My Ludum Dare 72-Hour Jams
LD#24 -- One Last Chance (Unity web & Android .apk)

::Post-Compo builds::
- (-o^_^)-o HUGMONSTER o-(^_^o-)
- Burning Platform (on Google Play)


Ludum Dare 26
Ludum Dare 25
Ludum Dare 24
Ludum Dare 23
Ludum Dare 22

recursive frog's Trophies

Quality Reviewer Award
Awarded by caranha
on December 23, 2011

recursive frog's Archive


Posted by (twitter: @RecursiveFrog)
Wednesday, December 19th, 2012 1:43 pm


This is simultaneously the game I am most proud of and the game I feel needs the absolute most work to make it truly acceptable as a “real” game. The concept : A completely voice-controlled audio game played through your Android phone. You call your special agents and guide them around the enemy compound, to fulfill their missions

— ^^ Video illustrating gameplay ^^ —

As you can see, you speak to your agents and they respond to your questions and commands … when they are understood.

But what about when they aren’t?

This happens a lot, I think. I’ve heard a lot of people say the recognition doesn’t work for them. Unfortunately, they can never tell me what they say, or what the phone hears. Do they have a strong accent? Are they in a room with a lot of conversation? Is there a song in the background? These things could all interfere with the speech recognition.

Therefore, I’m very strongly considering releasing an update to the game that includes analytics. I would want to store, send, and study the following:

1. Which agent is on the phone
2. Which location they are in
3. What the phone thinks you said
4. What command your speech interprets into

Since this would be part of debugging I feel that this would not be breaking the rules of the Jam to issue an update to do this… But do you agree? If I do incorporate analytics, would you have concerns about the amount of data sent costing you money on your data plan? Any other concerns?

If you want to give the current build a try, before I make any final decisions on analytics, here is the submission page.

Play “Agents” for Android


Posted by (twitter: @RecursiveFrog)
Monday, December 17th, 2012 3:20 pm

We’re approaching the target. Command?


Our jam entry is nearly complete. We had to cut a bit from the scenario, and I’m also concerned that I might have the only phone on the planet that won’t choke on all the audio we’re juggling, but we’ll see soon! We’re on target for a successful launch shortly.


Posted by (twitter: @RecursiveFrog)
Sunday, December 16th, 2012 4:14 pm

Because you shouldn’t have to remember the agent’s phone numbers, here’s a contacts list with all the numbers you need to know.

It also doubles as the credits screen. :p


frog.recurse(goat) throws IllegalArgumentException

Posted by (twitter: @RecursiveFrog)
Sunday, December 16th, 2012 4:02 pm

Believe it or not, this game in which the only UI is a phone dialer… gets a +1 Goat.



Posted by (twitter: @RecursiveFrog)
Sunday, December 16th, 2012 3:40 pm

What I love about the 72 hour team jam

You have some more time for polish and bug fixing. Also, people with strong skills in some areas and weaknesses in others can come together to make something super sweet!

What I don’t love about the 72 hour team jam

All you 48 hour guys are all wrapping up right about now and I won’t get a chance to really play your awesome stuff until tomorrow, and likewise you won’t be able to play mine until then either. Sadface.


Posted by (twitter: @RecursiveFrog)
Sunday, December 16th, 2012 2:12 pm

So, what is it like to play a game where you don’t see anything, but only know what you can hear?

Here’s a video with unpolished voiceover for Agent #2 for your enjoyment. Right now, all you really do is just wander the halls and listen to descriptions. You *can* ask to know your status or inventory but there isn’t voiceover for it yet.

Yes, the bug in the beginning of the video should be resolved now. :)


Posted by (twitter: @RecursiveFrog)
Sunday, December 16th, 2012 1:29 am

This game has VOICE ACTING.


I have been writing and recording VO for Agent #2, a sneaky, shifty thief, for the past 3 hours almost.

My sound partner is doing the part of Agent #1, a demolitions expert who is itching to blow something up, for almost the same time.

Considering that your only window into the game is their verbal descriptions of what the Agents see, and what you hear around them in the ambient sounds through the phone, this is all incredibly important. Aside from just a bare phone UI, there really are no graphics.

And, by the way, it’s proving incredibly trippy to hear the ambient sounds change as you direct your agents through the building. This is turning out so much better than I had imagined and we’re still only halfway!


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 6:06 pm

You say into your phone : “Agent 2, head to the sewer”

You hear from your phone : the sounds of a sewer when the Agent moves there

No screenshot obviously. :p


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 4:41 pm

You can ask the Agents questions now, such as “Where are you?” and they will tell you what room they are in.

Likewise, you can tell the Agents to go to an adjacent area, or even an area that isn’t adjacent but that they know how to find, and they will do that too. You can use relatively natural language for this as well. So when you speak phrases into your phone like:

“Go to the courtyard”
“Head over to the courtyard”
“Make your way to the courtyard”

All will basically just translate to MOVE COURT_YARD.



Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 3:21 pm

This is how what you say is interpreted by the phone, and in turn how it is delivered to your Agents.

I said “Get the blueprints”.


And next, the Agent will attempt to get blueprints, if those are in the room with him.


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 1:07 pm

Meet your nameless, faceless henchmen:


Agent #1

  • Demolitions expert
  • Tasked with destroying the weapon prototype
  • Is strong, but not stealthy

Agent #2

  • Expert at picking locks and stealing items
  • Expert at computer systems
  • Is stealthy, but not strong

Using only what you can hear from their surroundings and what they describe to you, can you guide them through this top secret weapons facility to wreck your opponent’s only hope for survival against your coming onslaught?


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 11:07 am

The floor plans for this game might need to be simplified a little. Don’t want to blow out the scope.



… wat.


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 3:11 am

It’s time for bed.

To recap here’s what we’ve created thusfar:

  • 75% of the game’s UI and Graphics
  • Got the speech recognition up and running (omg)
  • Settled on a plot for the game
  • Started drawing up a world map

And now, off to the bed!


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 2:33 am

What does this:


Have to do with this?


You won’t ever see those floorplans, or ever explore them personally.

That’s what you have Agents for.

Can you guide your Agents through this heavily guarded building from your command center?


Posted by (twitter: @RecursiveFrog)
Saturday, December 15th, 2012 1:12 am



Speech Recognition API — Good to go!

You’re gonna control this thing by talking to it.


Posted by (twitter: @RecursiveFrog)
Friday, December 14th, 2012 11:16 pm

And if all goes according to plan, here are some APIs I’m going to use.



Did I just get myself in *WAY* deeper than I bargained for? 😉


[cache: storing page]