ancient wisdom

Palm held a competition to promote their PDK. Ancient Frog was eligible, but (possibly because it had passed its peak before the competition started) didn't fare as well as I'd hoped.

On the other hand, there was a category for free apps, and for a while the number of entrants was smaller than the number of prizes on offer. This looked like a good opportunity to bang out some free apps - it meant I might win something, I could experiment with what genres of app generate the most interest, and I could have a bit of fun doing some rapid prototyping. All models of Pre and Pixi have OpenGL ES2.0 hardware, so I could go to town with shaders.

I gave myself no more than two days per app. The speed of development and the limited scale of what I expected to achieve meant that it was a proper hackfest. All the code went into one ugly C file. Magic numbers, hard-coded data, copy-and-paste coding - everything that shaves a few minutes off writing a bit of code at the expense of making it unmaintainable, bug-prone and difficult to build upon. I felt simultaneously dirty and liberated.

The first thing I made was Earth Now. It shows a 3D view of the Earth (rendered using NASA satellite imagery, including normal mapping from the height data and specular mapping for water / ice), and downloads the latest global cloud cover imagery to display on top. I thought it would be a quick app to ease myself in to things, but it ended up taking two days - most of which was locating a decent source of cloud imagery, and writing my own http code to get at it through its content network.

The next one was a sandpit app called Liquid Planet. It combined a bunch of tricks that I found useful in museum interactives - GPU-based interactive water and earth, with a bathymetric model of the Earth that you can flood and drain. It's quite fun drowning continents, or seeing just how much land there was during the last ice age (New Zealand was huge!). This one took less than a day, but was unfortunately a bit too demanding of the phone's GPU. (I'm used to doing this sort of thing on high-end PC gaming cards, and while it's impressive how much you can do in a shader on a Pre, it can't really handle the number of textures I wanted).

Then, a music app called Tonal Pool. I've traditionally been very bad at handling sound in my games, generally seeing it as the area responsible for the majority of the game's data footprint, the majority of the annoying configy crashy bugs, and the least important part of the experience. (Throughout my time in the games industry, I'd hear people repeatedly argue that it's the most important part of the experience, but really - try playing a game with the speakers turned off. Oh, you already do! Now try playing with the screen turned off.)

Anyway, it's time I made my peace with that particular sense, so I decided to make an app that was all about the sound - a Tenori-On-style sequencer. Of course, I still didn't trust it to survive on sound alone, so it's framed in a nice little GPU-based interactive water pool, with ripples going off as each peg sounds. I was pleased with the result - one day of work, and I had several nice emails about it.

Banner came next. It was a bit of an experiment - I wanted to see if I could create a software version of those light sticks that you wave to spell out messages using persistence of vision. The screen can only handle 60fps, which is definitely at the low end of what you need to pull off the trick, and there's an accelerometer to help sort out synchronisation.

You draw your design on the screen with a rudimentary paint application, then start waving the phone. It automatically senses this, and starts flashing a line of dots to recreate your image in the air. It works, but only in a really dark room, and it takes some practice to get good results. After a day on it, I decided that it was never going to be good enough to meet the Free App Downloader's expectations, and rather than get slaughtered in the comments, I decided not to publish it. (Search the iPhone app store for similar apps if you want to read how people respond to basic physical limitations...)

Finally, my favourite: Ancient Pond. (Yes - I liked it so much I gave it the Ancient seal of quality). It's your basic fish pond app, using (ahem) that GPU water effect again, this time with a nice flocking algorithm running under it to make fish dart and swim about. It took a couple of days, and then another day or so updating it to add different themes and do some tweaking of the experience. I managed to keep the framerate up to an acceptable level (particularly on the Pixi, which, surprisingly, handles it better than the Pre) by strictly limiting the effects. So there's a theme which has shadows and refraction on the water, and a theme which instead has specular highlights and caustics.

Ancient Pond took off really well - there's clearly an interest in this sort of app, and I really enjoyed myself writing it. It's very tempting to turn it into a real app, cross-platform, with downloadable creatures and ornaments and so on. I'm still mulling over whether to put Blue Comb on hold and do this first.

In the meantime, it would be nice to add a few more features, but it's right up against the hack-it-now-and-never-look-back approach I used to create it. To do anything more to it at all would really mean tearing it down and starting again properly.

So, how did I fare in the competition?

Well, the results don't seem to be up yet, but I've had an email which confirms that my placement in the leaderboard was accurate. Ancient Frog snagged one of the $1000 HP store credit prizes, as did Ancient Pond and Tonal Pool. Both of them hovered tantalisingly in the $10,000 actual money prize category, but were finally beaten out. Earth Now and the free version of Ancient Frog were also in the running until the last few days of the competition. Liquid Planet fizzled, and I ended up pulling it from the store.

I had a lot of fun, and I have a nice little purse which I have to spend on fun things, so that's great. And since the HP store sells Canon lenses, and I live for Canon lenses, I'm a happy man.

And here I am again, shooting my mouth off to Intel:

http://appdeveloper.intel.com/en-us/article/iphone-appup-porting-ancient-frog

Ancient Frog currently runs on all of the devices in this picture, and a few more that happened not to be there when I was playing with my camera.

The iPad port is out and failing to set the world alight. It did pretty well in its first week, getting featured by Apple. But featured on the iPad in launch week turns out to fall a long way short of featuring on the iPhone, and when the week was up, it didn't have any momentum.

The Palm Prē version is out too. I really enjoyed doing that one - I discovered at GDC that Palm was opening up native development, got a Prē straight away, and had something running in a couple of days. It's essentially the Windows build in its iPhone-sized debugging configuration, with the OpenGL ES code and textures from the iPhone version. Wonderfully painless, and with the crisper screen, and the river-polished-stone feel of the Prē, it's beautiful. My favourite version so far. It takes a while for the sales figures to come through from Palm, so I don't really know how it's doing yet - but it doesn't need to do anything spectacular to justify the effort.

The Android port is still rumbling along. I have something that runs well on the Nexus One, although it lacks sound, there are some situations where it fails to recreate textures correctly, and the levels all need redoing (again). (This time they need to be higher res and aspect-ratio agnostic versions of the original iPhone levels - the iPad versions only make sense on a physically larger screen). The textures also can't be compressed, because there's no standard GPU on Android devices, and there's a silly restriction on how much data you can fit in internal storage. So lots of bitty work to get it to run reliably on a decent cross section of the various devices out there, and still no way to charge money for it from New Zealand. I keep picking away at it because at some point it'll probably be worth the pain to nail all of the remaining problems.

The desktop version, OS-X and Windows. I keep flip-flopping on this. I'm just not sure what sort of market there is for a game like Ancient Frog on the desktop. The App Store has lowered everyone's price expectations, and I don't think I'd be able to sell nearly enough copies at $5-ish to be worth the support hassles of getting it running on decade-old malware-ridden boxes.

The netbook version. This is basically the desktop version, but with Intel handling some of the config testing woes, and the possibility of prizes in their competition, I'm tipped in favour of it.

With all of this stuff, Ancient Frog continues to take up a couple of days a week of my time. The rest of the time I'm working on two new game prototypes, hoping to get one of them to a critical point where I can announce it and go all in on it.

You may have missed it - the news slipped out pretty quietly - but Apple has announced a new product line.

The iPad is essentially a scaled up iPod touch, so naturally I'm interested in what it means for Ancient Frog.

According to the announcement (and my experiments in the simulator), iPhone apps work on the iPad straight out of the box, with no changes needed. However, since the screen is higher resolution (1024x768 instead of 320x480), and a different aspect ratio (4x3 instead of 2x3), such apps won't look their best. Below are a couple of mockups of what you'd see if you had an iPad right now and ran Ancient Frog on it. (All of the pictures are shrunk down to fit on your screen - click on them to see them full res.)

The first mode just runs the iPhone app at its original resolution, giving you a little window in the middle of the screen:

It also gives you the option to run it "2x" - doubling the size of each pixel (in each direction), so the app uses more of the iPad's screen (640x960 of it). This makes it bigger, but either blurry or blocky depending on whether or not the device is going to filter the image when it scales up. I suspect they'll go for blocky, which is what I've simulated here.

Now, neither of these options is particularly elegant or attractive. With a bit of fiddling around though, I can make the current iPhone application recognise the iPad and behave more like a native application on that platform. What I've done here is run it at 768x1024, but allowing it to letterbox slightly to retain the original aspect ratio (luckily the ragged border gives me a neat way to bring the edges in a bit, as well as a bit of room to lose some pixels top and bottom). This already looks way better than the previous shot - lots of elements are still blurry, but things that appear at varying scales in the game are already at a higher resolution. This means the text, the daisy and the particle effects are all crisp, which makes the whole thing seem higher resolution.

So that's good - I can put a bit of work into the current app, and release a free update that makes it play nicely on the iPad. But it's still blurry - what would be involved in making it look great on the new hardware?

That's where it gets a bit tricky. The work involved suddenly goes up by an order of magnitude. Every level has to be reworked, and I have to handle the original aspect ratio, as well as two new ones for the iPad (for portrait and landscape). The download size also goes up - currently I fit just under the 10MB limit that allows people to download over their phone. With the background textures doubled up, I'm over 20MB and start to lose impulse buys (as well as bloating out the app for everyone, regardless of whether they have an iPad).

So my plan is to leave it at that - an incremental upgrade to make the experience worthwhile on the iPad - and offer a separate, HD, version for iPad only. This will use the reworked levels I've been beavering away at for the last few months, and make use of its groovy auto-aspect-ratio handling (the desktop version can be stretched to all sorts of whacky resolutions without breaking the effect).

And here's the result:

It's work in progress (the lighting has yet to be graded correctly, and the particles are missing), but you get the idea. The full screen is used, the frog is back down to a more appropriate size for human fingers to prod at, and there's some nice little environmental stuff going on for you to play with.

It would be nice if there were a way to offer a cheap upgrade path for people who already have the iPhone version, and I'm looking at whether there's a clever way to do that.

I can't wait to get my hands on the physical hardware. I think there's going to be some very interesting times for gaming in the next few years.

My thrashing around trying to pin down the details of project Blue Comb continues. I have decided on an art style, which is helping to frame the rest of it. Ancient Frog started with a very simple and clear premise, but this new project is just a collection of obsessions which I haven't yet managed to glue together. I suspect something will need to get cut before I'm able to finish it.

I just received an email from Jef Armstrong, whose game Mondrian was recently released. Jef emailed me last year, shortly after Ancient Frog was released, saying words to the effect that he liked my puzzle game, he was writing his own puzzle game, so let's talk about puzzle games. Flicking back through the emails, there's some blogworthy stuff about how I created the levels in Ancient Frog:

Creating the puzzles was one of the big challenges with Ancient Frog. I've been thinking about writing a blog article about it, because it's something I personally find fascinating. A large proportion of all the code written for Frog (possibly the majority - I should check) went into the tools for creating the puzzles. However, I'm a bit wary about talking too much about it in public because I suspect that some people would be disappointed to learn that it's not all lovingly hand crafted by some sort of puzzle making master craftsman.

Seems a bit silly now. Let's hear the details!

The most important part of the toolset is the solver. This has two levels - a quick pass to determine whether a particular layout has any solution at all, and a slow pass to exhaustively determine the optimum solution. (In this context, 'quick' means anything up to 15 seconds, and 'slow' can be up to 6 hours). There's nothing clever about the way it works. It uses brute force and a fast computer to walk down the tree of possible moves. There's some interesting optimisation, and stuff like exploiting symmetry, early-out, loop detection and so on, but essentially it uses the approach that computer scientists say is impossible. But that's because it's their job to come up with general solutions which work for the worst cases in the biggest data sets. I'm able to limit the problem space to something which is achievable by brute force, and just skip puzzles which are taking too long to check.

So the process for generating a puzzle goes like this:

* I choose a start point and an end point, and a set of pegs from which it can choose its route (this lets me put a particular move into a puzzle, but often I'll just leave it to choose from all pegs)

* I press the magic button

* It goes through a loop where it removes a random peg and tests whether a solution exists. If there is no solution, it replaces the peg and flags it as required. This process is repeated until no peg can be removed without creating an impossible puzzle.

* It checks the result against all the previously generated puzzles to ensure it's not a duplicate (taking into account simple translation, rotation and mirroring)

* If I like the look of the result, I press magic button #2 and generate the optimum solution for it

* I quickly step through the solution to see if I like the look of it. I tend to reject puzzles that require too much walking backwards, or that push the frog's head off the side of the screen or have the legs passing over the fly.

I like to generate a whole bunch in one sitting, then try to put them in order based on how difficult I think they look, and how difficult the solver reckoned they were (that's a bit of code that needs some work by the way - some of my rankings are quite a bit off). Then I leave them for a day or so while I get on with other stuff, so I can play them through without knowing how they're solved. In my playthrough, I log how long it takes me to solve, how many dead ends I go down, how many undos etc, all of which information also goes into the difficulty ranking. Each time I play through, I'll generally shuffle levels around a bit to improve the progression, but I also like having runs of very similar layouts with increasingly difficult solutions.

The tools also include stuff for applying backgrounds, frog species, fine-tuning of board position, rotation & translation of the whole puzzle and so on.

Interesting stuff!

Happy new year.

A couple of days ago I put Ancient Frog on sale ($1.99! Down from $4.99! This crazy offer cannot last!) The last time it went on sale was back in May (with rather disappointing results), and I thought it was worth trying again. The App Store is constantly evolving, after all.

I was impatient to see what effect the price drop was having, so I started stat mining. Previously, I've tended only to look at the bigger markets - originally the overwhelming majority of my sales came from the US and UK, later adding Germany, Canada and Australia. Over time, though, the distribution has started to flatten out more. The US is still the biggest of course, but it now accounts for under 50% of my sales, and Japan is up to nearly 10%. With that in mind, I decided to just grab everything - every ranking from every region - and see how I was doing.

One figure jumped out at me immediately: in one of the rows, Ancient Frog was ranked at #1. The number 1 paid app! Woohoo! Still got it, 10 months after release! Number 1... in the Family Games category... in Paraguay. But who cares about the details? That's got to be worth a screenshot:

In fact, digging further, I saw that in the Paraguay App Store I was not just the #1 paid app in Family Games, but the #2 Puzzle Game, #2 Game overall, and the #4 paid app in the entire region. I owned Paraguay! This is how fortunes are made, right?

Well, it turns out that to reach these dizzyingly high rankings, you only need to sell one copy. I had a trawl through my stats, and precisely one Paraguayan has bought Ancient Frog. (Thanks, whoever you are! If I'm ever in Paraguay I'll buy you a drink.)

Interesting, but probably not that surprising. But then I had a thought - just how much ranking does one purchase get you in all of the other regions? This price cut has led to a very broad but very shallow spread of new sales, and in a number of regions I've just made my first sale. In the following table, I've listed all of the regions where I've sold only one copy in at least the last two weeks. The purchase generally took place in the last couple of days, but I've listed some where I'm still ranked despite an interval of several weeks. The positions noted are the highest that I saw - these can drop off quite quickly (because we're necessarily dealing with regions with very granular figures).

So it's all a bit haphazard and imprecise, but I think it's interesting for all that.

App Store ranking achievable with a single sale:

Country        Family Puzzle Games Apps
_______________________________________
Paraguay       1      2      2     4
Czech Republic 18     39
Slovakia (1)   20     36     50    94
Finland        28     71
Hong Kong      31     60
Poland         41     83
Portugal (2)   42     91
UAE            54     92
Luxembourg (3) 59
China          67
Austria        72 
Kuwait         75
Singapore      84
Turkey         90
Thailand       92
Switzerland    98

1 I saw it at #4 in Family in Slovakia, but when I took this set of figures a day later, it had dropped to #20
2 Copy sold on Nov 21
3 Copy sold on Nov 28

Now you might say that this isn't terribly useful or interesting information - obviously regions with low sales won't take much to break into their rankings. You might say that, but have you got the number 1 paid app in Paraguay (Games/Family subcategory)?

No you haven't.

I have.

Back from an extended visit to the UK. It was good catching up with all the people I haven't seen since I moved to NZ, and showing off the son and heir.

I'm rapidly approaching the end of some paying work, and when it's finished I'm planning to go full time on Ancient Workshop. It's been really annoying this year, watching opportunities drifting past for my own games while I work on somebody else's stuff. Definitely time to give it the attention it deserves.

Meanwhile, here's Ancient Frog on a list of "10 of the Most Beautiful iPhone & iPod Touch Apps" - makes my day.

I'm continuing to eke out Ancient Frog's tail with little incremental bumps. 1.10 adds another level. I'll probably keep going at this rate on it for a while yet, giving me some space to get the last of the museum work out of the way and finish up at least one other platform.

I'm champing at the bit to get going on something brand new and shiny, but while there's still life in the Frog, it makes sense to feed it.

Removing the free version had no immediate effect. Sales trundled on unchanged for about a week, then started dropping - which they were probably due to do anyway. I put LE back on the store, and have released the 1.09 update.

1.09 contains a brand new level (albeit in an existing habitat), as well as the latest version of the Ancient Engine.

1.07 was so long in Apple QA that I'd finished a nice clutch of UI improvements by the time it came out. I decided to push it through immediately, to see what would happen if I got out an update while the boost in visibility was still in effect from the previous one - whether there'd be any advantage to it. As it turned out, it got buried in the mass of updates they got through while clearing the decks for the 4th of July (apparently some sort of cause for celebration in the US), so it didn't really do me any good.

1.08 fixes some nasty flickering that could happen when turning the page in the UI, improves performance (or at least battery life), and adds better feedback to the buttons. The buttons in the game started life on a mouse-based machine, and they didn't give terribly good feedback when deprived of a mouse hover condition. This meant you could be left wondering whether you'd actually pressed them or whether the game had hung or what. Now there's a glow effect that pops up the moment they activate, and stays there if the game has to lurch for a moment loading data. It's something that really should have been in from the start. Better late than never.

I haven't started on the next update, but there will be one. It seems to be the most successful way to keep the game visible. Downloads of Ancient Frog LE show a step up followed by mathematically perfect logarithmic decay after an update. (Downloads of Ancient Frog itself are low enough that the noise spoils the perfection of the curve.)