The Real Pandora Niche

[MWeston]

Correct me if I'm wrong, but didn't someone from OPT have to finish the nubs firmware? That means you guys own the source code, correct? Would it be possible to take a look at that source? If so, an unofficial software hack may be possible, assuming there's anything useful in the source code that could be used.

The actual algorithms for calibrating and reading the nubs is proprietary. They sent me an object file for that and an open source project package for the rest. The HAL, power management and I2C communication stuff is open source since it is basically all pulled right out of Atmel's free library of code. Because of how the code is, I have no access to the calibration data and even if I did, I don't know how useful it would be because the nubs are always self calibrating.

The other problem for adding any new features is that these chips have 8KB of flash ROM and it's almost full so there isn't much I can do unless we buy bigger chips (making the first 4000 customers left in the dark). I'm not sure if the 16KB's come in the same sized package but I haven't looked. Right now, reflashing the chips requires access to the custom header on the front side of the PCB and using a custom 'bed of nails' programmer I built (which got all smushed in shipping to England damn it!).

Also, Craig was referring to using regular carbon resistance based nubs like the PSP which they might have some copyright on. The Pandora nubs use a different technology that I probably won't volunteer information on (they haven't told me if I even have permission to do so yet and I did ask) but if someone rips it apart, you will see how it differs.

 

[cosurgi]

note to myself to look up sometime in the source code. An idea for nice behaviour of nub. fractional calculus integration of nub position "velocity" to mouse. derivation stage parameter. I can explain later. Now just writing it here to not forget

 

[MDave]

Oh well, guess we just gotta twiddle those nubs each time we want to play with Pandora

 

[urjaman]

The actual algorithms for calibrating and reading the nubs is proprietary. They sent me an object file for that and an open source project package for the rest. The HAL, power management and I2C communication stuff is open source since it is basically all pulled right out of Atmel's free library of code. Because of how the code is, I have no access to the calibration data and even if I did, I don't know how useful it would be because the nubs are always self calibrating.

Can you describe the pinout of the header on the PCB?

Are you allowed to provide us the firmware source and the object blob?

If you cant, are the AVRs "read protect" bits set or not (eg. can I read out the flash with the programming interface)?

I'd love to atleast attempt to squeeze in a boot loader so the AVRs could be reprogrammed over I2C.

EDIT:
And yeah i dug out the best picture of a top side of the pandora board i could find (it was r4 but i think for this it'll do). I see a 2x3 programming header (for both nubs) that looks suspiciously familiar (for an AVR geek) .
Short Question: Does it match Atmels 6-pin layout?

 

[cosurgi]

sounds a bit like mouse acceleration is enabled

my idea is just about that. Usually in differential calculus you can have just a velocity (1st derivative of position) or acceleration (2nd derivative of position). The nub position can be either "velocity" so you integrate it ONCE to get position, or it can be an "acceleration" so you integrate it "TWICE". Fractional calculus allows you to integrate 1.23 times. So then the nub position will be a something between a velocity and acceleration (in case of 1.23 it would be closer to velocity). Later I'll churn out the formulas for you. Then we will just precalculate the values and make a lookup table. I'm basically curious if this will alloow comfortable usage of nubs.

 

[hobbyman II]

this is why I hate all this "IP" cr*p, you want to do the simplest thing and end up jumping through hoops and re-inventing the wheel, all for something that's so bloody obvious it should NEVER be given a patent, pity the nubs can't just give a plain and simple X,Y value, but I guess some donkeys a*se patented X,Y values for HID controllers or something, give me strength :angry:

 

[SomeGuy99]

this is why I hate all this "IP" cr*p, you want to do the simplest thing and end up jumping through hoops and re-inventing the wheel, all for something that's so bloody obvious it should NEVER be given a patent, pity the nubs can't just give a plain and simple X,Y value, but I guess some donkeys a*se patented X,Y values for HID controllers or something, give me strength :angry:

Just think if somebody did patent the wheel...

 

[Tom`]

this is why I hate all this "IP" cr*p, you want to do the simplest thing and end up jumping through hoops and re-inventing the wheel, all for something that's so bloody obvious it should NEVER be given a patent, pity the nubs can't just give a plain and simple X,Y value, but I guess some donkeys a*se patented X,Y values for HID controllers or something, give me strength :angry:

Just think if somebody did patent the wheel...

http://news.bbc.co.uk/2/hi/asia-pacific/1418165.stm

 

[Morn]

Actually, some Aussie did:
http://www.newscientist.com/article/dn965-wheel-patented-in-australia.html

-- edit --
h34r:

 

[hobbyman II]

ok! I WAS off the sauce, but seeing that (even if it WAS making fun of the law) has driven me back to drink, move over Craig, and pass me a glass will ya? first one to spot five pink Elephants gets to patent water (a mythical substance some people claim to be able to drink :rolleyes: ).