I’m not sure if what people want actually makes sense. These upgrade requests already exist in the market!
They’ll always want more ram/cpu/periferals, and bit-by-bit it’s going to start looking more like a Nintendo Switch.
Upgrading within the AVR line doesn’t make sense spec+pricewise.
Going to ARM would kill the 32u4 Arduboy (the same way the META deprecated the Gamebuino “classic”) and there’s a plethora of other devices already in the market throughout the power range:
Low: Gamebuino META, Pokitto, TinyCircuits, BBC Micro.
Mid: Raspberry/Orange/Banana/etc Pi
High: Android and iOS.
The restrictions of the 32u4 with a 1-bit screen are what make the Arduboy interesting to me, so an Arduboy 2 should be strictly binary-compatible, both ways. Maybe a bigger screen, but still based on the SSD1306 (is there such a thing?). Maybe an SD card with a bootloader that can load games from that. Maybe a variant with a vertical screen.
Looking at the polls you can see that people have opted for 2-bit colour rather than 4-bit colour, there has been no interest in the microphone or photosensor and most people wanted roughly the same speaker.
I think most people want something that’s a bit more powerful, but not too much more powerful. Probably something closer to the original Gameboy in terms of power.
I think the Gamebuino Meta would be a better analogy since the DS and PSP aren’t intentionally reprogrammable (you have to homebrew if you want to make games for them, which is of questionable legality).
A 2bit screen would be nice since people wouldn’t have to rely on flickering for a third tone.
I feel like the chip itself is fine power wise. Like I talked about in the ardumon thread, the gameboy carts have a lot more memory in some cases. The arduboy having a capacity of ~3% of the largest carts (what pokemon red/blue use). I didn’t feel like the biggest bottleneck in Arduventure was the capability of the chip or screen size/color depth. It’s clear where features had to be cut down to squeeze every bit of memory. I haven’t compiled Arduventure so I can’t comment on the ram usage in comparison to a gameboy, but just having more PROGMEM would be a huge boost to the possible scope of any game.
The hardware being so basic is what appealed to me so that’s just my bias at least. I think backwards compatibility would be nice, as @FManga suggested. New games being playable on the old arduboy would be nice too assuming they didn’t use the added memory in my idea’s case.
An Arduboy with a 2.5in 128x128 px 2 bit monochrome oled/lcd screen, more ram and eprom, if necessary with a SD card, a start and select button, and an SD card slot will make me happy. Then it will be a lot more close to Gameboy both look and power wise, guessing that the hypothetical future Arduboy is still designed to look like a Gameboy, but it’d be also nice for it to rock it’s own look.
A decent chunk of the time that went into Dark & Under was going around cutting the size down or deciding which features to drop from the original design because of size constraints.
In fact if it hadn’t been an issue, I probably wouldn’t have been called in to help in the first place.
It also hasn’t stopped my own games from progressing slowly or features creeping in.
While I still support the 2-bit screen wish, I agree with this.
The bottleneck is almost always the progmem.
Though I think if there was an SD card that could be read from then most games near the progmem cap could load levels from the SD card, in which case RAM would then become the bigger issue.
(E.g. Arduventure could load its maps from the SD card, which would save a fair chunk of memory.)
Why not just step up from Arduino Leonardo specs to Arduino Mega?
256KB Flash / 8KB RAM, the assembly code portions of the library remain compatible, and you retain the Arduboy’s 8-bit spirit!
Seems like all of this is moot though - the decision to jump to Arduino Zero specs has already been made, but I have to agree it makes me sad somehow that the Arduboy / Pokitto / META will all be essentially the exact same device, but with 3 different screen variations…
Remains compatible with the huge number of current Arduboy games at the C++ source code and library level, requiring just a re-compile for the majority of games.
Doesn’t add a lot of new “whiz-bang” features that make the overall unit capabilities difficult to grasp and more complicated for “noob” programmers to work with.
Keeps the same credit card size. (Preferably the exact same case, with minor mods for an SD card slot and/or other decided on additions.)
Costs less than the two mentioned in the quote and any other similar ones.
You have to consider the momentum that the Arduboy has, in terms of numbers and game availability. In addition to official Arduboys shipped, look at the number of Chinese “clones” available as a testament to its popularity. There are literally hundreds of games and other sketches available for the Arduboy (with over 20 more added just in the last few weeks due to the game jam). A substantial number of those are well done and very enjoyable to play.
If the next generation Arduboy strays substantially from the current one, much of that momentum will be lost and, as @uXe says, it will just be another new, different device in a sea of competitors. (Honestly, I’m sceptical that the Pokitto or META will ever gain the popularity that the current Arduboy has.)
So, as my philosophy has always been: Give developers a bit more processor power and memory resources, so they’re not pulling their hair out wasting time on these, but keep it as backwards compatible and inexpensive as possible. It should remain slotted as a relatively simple, easy and fun beginners/learning platform that can none the less produce decent, enjoyable games. My Arduboy Z prototype embodies this philosophy.
As for the popular desire to have more than just a monochrome display, IMHO it would have to remain compatible with the commands, resolution and speed of the current display (at least at the library level) when run in 1 bit mode, so as to remain backwards compatible with current games. Its size should be a consideration in terms of maintaining the unit’s credit card size. And, its cost shouldn’t impact the desire for a low cost unit. I challenge anyone to find such a beast.
Okay, I closed the polls, as it’s pretty clear what we want for the next generation Arduboy.
Also I agree with most of these statements.
That the Arduboy is unique with it’s “credit card” gaming system.
I would like the most if they added SD card support, so I can play multiple games without a PC (if I get bored with the current one).
And a connection between 2 arduboys (like the Gameboy’s link cable) would be pretty cool for the system.
I haven’t looked into this much, but it’s possible that the extra code required just to support the SD card might limit memory available for games, such that a lot of SD data loading (which is relatively slow) would be required for many games wishing to make use of it.
Also, if you had an SD slot, you would probably want to use it to load different games (as well as being used by a game to load/store data). This would best be done by adding such capability to the boot loader or library, but expanding the size of these on the 32u4, while remaining able to load large existing Arduboy games, could present problems (though it’s possible that a version of @Mr.Blinky’s optimised bootloader may be able to squeeze something into the 4K used by the Arduboy/Leonardo bootloader). By switching to a new processor with a fair amount more space for the bootloader and sketches, this problem is probably eliminated.