Windows: I have written a Windows program (source) based on John's Mac software that allows you to download pictures, clear the camera's memory so that you can use it again and again, and perform memory and register dumps.To be written. If you're here, you probably already know what these cameras are and where to get them. Picture quality is a bit lacking, but acceptable for Web images and the like, and certainly not bad for $11.
Note that there is a similar PureDigital single-use camera available from participating Walgreens stores. This camera may be identical, similar, or even completely different internally from the Dakota camera available from Ritz/Wolf Camera. We have not yet tested these, and thus do not know if any of these files or information are applicable to the Walgreens camera.
Pin 10: Ground (Black)
Pin 9: Data - (White)
Pin 8: Data + (Green)
Pin 6: +5V (Red)
The colors listed are the standard color-codings of the appropriate USB cable wires. The pins are numbered by the "1" and "10" on either side of the edge connector. You can either solder the wires of a USB cable directly to these edge pins, (better) solder your own connector into the camera and a matching connector on the end of your USB cable, or (best) find a connector that already fits the edge connector with no soldering required. I have not found such a connector, but have heard that a Palm hotsync cable will fit with minor modifications. I have tried what I think is a Palm-to-serial cable (found it somewhere), but it did not work.
NB: Don't get ripped off buying a USB cable (only to cut one end off of it) at well-known computer and appliance outlets such as my fictional store, BestCircuitDepotUSA. (Particularly avoid any place that also sells USB printers and scanners; these will artificially inflate the price of the cable to make up for slim margins on the printers, which are intentionally sold cable-less.) You will spend substantially more for the cable (avg. cost $20) than you did for the camera. Try cablesforless.com (avg. cost $2) instead.
Macintosh: Maushammer has written a Mac program to download pictures from the cam. This program allows you to download pictures, clear the camera's memory so you can use it again and again, and supports advanced developer/hardware hacker features such as raw memory and register dumps.
Windows: I have written a Windows program (source) based on John's Mac software that allows you to download pictures, clear the camera's memory so that you can use it again and again, and perform memory and register dumps.
Unix/Linux: gPhoto, a Linux program supporting over 400 makes of digital camera, is expected to soon have full support for the Dakota camera. Until the new code is merged, you can grab updated copies of the relavent gPhoto files from John's site.
Despite claims heard that the camera is 2 megapixels, it is actually 1.3MP, and produces images with a size of 1280x960 pixels and average file size of about 400kb (although this varies greatly depending on the subject matter). Like many digital cameras, the pictures are somewhat darker than one would expect, and often can benefit from some lightening or gamma-correction tweaks from your preferred photo editing software.
Don't expect a lot of detail or clarity from your pictures (a paraphrase that comes to mind is "Unclear at any distance"); do expect a minor JPEG mess as the cam tries to guarantee that 25 pictures will fit into 16MB (or slightly less) of flash memory.
Click here to see some sample images from the camera.
Since I have noticed substantial motion blurs in some of my pictures, I figured I would measure the exposure length experimentally. The camera's "shutter speed" appears to be about 70-75 ms, or about 1/13 of a second both in medium light and extreme darkness; lighting level does not seem to have an effect. (These pictures are of a 'scope at 0.02sec/div; to my surprise, this actually turned out to be an acceptable speed indicator, relatively unaffected by residual phosphor glow.)
For reference, many professional photographers will insist on the use of a tripod at speeds slower than 1/60 sec. To avoid introducing more blur than usual, make sure the camera is held as still as possible.
The lens of the Dakota/Walgreens cameras is adjusted at the factory, then secured in place with a blob of clear glue. A metal spring between the lens and what it screws into also helps hold it in adjustment. It is probably adjusted for 'infinity' or near enough to it, but if you cut/dig/chip away the glue, you can rotate the lens to adjust the camera's focal length. Turn clockwise to focus far, and counterclockwise to focus near--up to within an inch or so of the lens! Note that if you turn the lens counterclockwise (unscrew it) far enough, it will come off in your hand and you'll have to screw it back in. With the lens completely removed like this, you get a nifty view of the CMOS image sensor, but will not be able to take any discernible pictures.
By performing this adjustment, it is possible to take crystal-clear pictures at a wide range of distances. However, you will then be limited to taking pictures only at that distance without adjusting the lens again, as the sweet spot is very narrow. What I have done is pick up two Dakota cameras, leave one focused as-is (infinity), and adjust the other to about 10 inches for close-up work.
Check out John's flash-storage-format page for an in-depth look at the organization of memory and files. This steps through a by-hand recovery of files, and even an attempt at recovering a previously deleted file. While that was being written, I was also hard at work on some quicker n dirtier documentation of FAT12 and how to read files from it.
While the filesystem is standard enough, the physical locations of data on the chip are all over the place. Luckily, this is also standard; this randomish placement of data (known as wear-leveling) ensures that each block of memory on the chip has data written to it an equal number of times (or as equal as possible, anyway). This is because Flash memory bits can only be written a finite number of times--although each bit is rated to be reliably written over a million times, it makes no sense to use the same bits over and over, wearing them out, while others sit relatively unused. John's site also covers remapping the wear-leveled data.
I apologize if I have forgotten anyone here - quite a few people and sites have been involved, and/or contributed in some way to the success of this project; almost too many to keep track of.
