3 rules of iphone development: 1) Mobile games *MUST* be able to be played silently. I.e. if you have speech for all your characters, make subtitles! it makes it so much better on us gamers... 2) don't rely on the UDID only for online leaderboards. I'm on my 4th ipotch in as many months and its frustrating as heck in some. 3) don't use the accelerometer just because it's there! seriously! this cannot be repeated enough. In fact I would venture to say "don't use the accelerometer except in specific instances (like Rolando, or a flight or driving game, or as an OPTIONAL scheme) because it's poor quality and noisy, which means you need to de-noise it by averaging samples over time which introduces lag and reduces playability, and really Apple should just put two-three accelerometer units in there and average the inputs of them in real-time to get something much more reliable and low-latency" and "the accelerometer is not disability-friendly" but that's all been discussed here I'm sure? caveat to #3: if you're VERY smart, you can do what Clinometer did, and use "dynamic sensitivity" to help with the issues of the accelerometer, (when values start changing big, reduce latency and go with it, because, say, the user is BANKING HARD and a little noise won't hurt as much, then lengthen the average stream when they get small again) but it's HARD, you have to sample at much more than the 26fps or whatever the ipod is locked to to make it work for a game (I'd reccomend 100hz, or around 4 times for each frame), and still, it's not disabled-person-friendly
Using a low-pass filter is better because it's slightly faster to implement and doesn't have lag. But generally I agree. --Eric
This does have its own disadvantages, and actually does provably introduce (some) latency. (I.e., a low-frequency trend has to develop before it will "pass," which takes more than one sample, which means latency). Perhaps dynamically setting the lowpass filter and disabling it over wildly-changing inputs would be a better solution...a hybrid approach. But still, the accelerometer is crap in the ipotch. *shrug* yes I will call it an ipotch. developers' exclusive use of the accelerometer with no other options in games where it has no place whatsoever disgusts me. You stop doing that, I'll stop calling it an ipotch. Deal?
Some good points--I've added them to my list here: http://forums.toucharcade.com/showthread.php?t=7015 As for the accelerometer, I've found it can be VERY precise (from the user's perspective) and the lag need not hamper game play. Dr. Awesome is the example that really sold me on the accelerometer for my game (not that I've started on it yet... but I plan to use tilt control when I do). The TiltMeter 2-axis level also impresses me with its fine responsiveness. Part of this of course is game design--and naturally, when designing tilt and touch-based games, you want to play to the strengths of tilt and touch! But regarding accessibility in general: I think games should consider offering multiple control methods when practical. Some people will be able to use tilt who can't reliably touch small targets. Others can touch but not tilt. Sadly, most games seem to really only have one control scheme that makes sense. So here's to the development of add-on hardware using the new iPhone OS 3! If a couple of good, accessibility-friendly control devices emerge and become popular, then I'd hope to see games widely support them. It would be cool to have one "standard" connection/signal method for one or two directional sticks/buttons/whatever and some extra buttons--and then ANY device following that "standard" would be equally compatible with games that understand the standard. I hope accessory makers get together on this.
