Direct mp3 playback (using the i300's RAM, CPU, and speaker) isn't really practical. The biggest limiting factor is the RAM. At 128kbps (the minimum bitrate that can be considered good fidelity), a single pop song would take 3-5 megs of RAM, leaving enough room in the i300 for maybe 2 songs (4 if you totally sacrificed audio quality).
If enough i300s get sold, there IS a practical alternative I can think of.
Imagine an mp3 player with its own memory stick slot (or 128mb of non-expandable ram at the BARE minimum), mp3 codec (the chip that handles the decompression), DAC (the chip that turns the bitstream into an analog audio signal), and stereo amp. This hypothetical device would be the same width and depth as the i300, and about an inch high. It would firmly clip on to the i300's hotsync socket, and be molded to fit the i300's contours perfectly and essentially become a lower extension of the phone. It might or might not incorporate its own lithium power cells, depending upon whether power is available from the hotsync socket and how much power the device would suck down. Although it would be self-contained insofar as mp3 playback goes, it would use the i300 as its control interface (song selection, volume control, etc.) and, as mentioned earlier, possibly as its power source. Oh, it would also have a passthru hotsync port on the bottom and (if it includes its own lithium cells) passively charge itself from the hotsync base (but ONLY after the phone itself were fully charged).
The cost? Probably the same as a standalone mp3 player (omitting the control buttons and display won't save much money, and the i300-specific nature would definitely drive up the cost a bit by limiting its potential market to i300 owners), but with a vastly better user interface (via the i300's color screen) than other standalone mp3 players.
A -MAJOR- value-adding enhancement, assuming this hypothetical mp3 player were to use memory sticks and the i300 could actually push the Dragonball's serial UART close to its theoretical max speed of 460kbps would be a mechanism for it to serve double duty as a near-line storage device -- holding large applications (Serious Sam, Rayman, etc.) for semi-casual use (with up to a minute or so of transfer and setup time), AvantGo data, images, etc.
Obstacles?
* Cost. It would be REALLY hard to justify paying more than $200 for mp3 playback alone, or $250 for mp3 playback + memory expansion.
* Speed. The Dragonball's primary serial UART has a theoretical top speed of 460kbps, but it's possible that something between the pins on the Dragonball and the hotsync interface itself might limit it to 115kbps (or maybe 230kbps). Like what? Well, the phone's CDMA chipset and the Palm's Dragonball chipset both share the same serial pins on the hotsync interface (otherwise, it wouldn't be possible to use the phone as a passive wireless modem). If both chipsets passively share the interface pins (by physically wiring both chipsets directly to the same pins with the expectation that only one at a time will try to access them), it's not likely to be a problem. However, it's quite possible that only the CDMA chipset is connected to the i300's hotsync port, and the Dragonball itself connects to the CDMA chipset and relies upon it as an intermediary interface to the outside world.
* Software support. Getting the i300 to recognize and use external storage is no minor feat. OS4 could handle it without problems, but the i300 is stuck wtih 3.52. So whomever were to market such a product would either have to license enough of OS4 to handle the external ram and come up with the software hooks to install it like one HELL of a Hackmaster Hack, or write the interface routines from scratch... hopefully copying Palm's interface specs closely enough to permit a naive application to think it was dealing with OS4 and work accordingly. Regardless of whether the major expense lies with licensing fees or development costs, it wouldn't be cheap.
Maybe we'll be able to find such a device under the tree next Christmas?
