The IDs are basically GUIDs - they are unique to each stream, and are currently generated by computing an MD5 hash from the time in milliseconds from the epoch - this would allow some server synchronization sometime in the future. If you manufacture them yourself, then you must insure that they meet the same requirements.

if they are symchonized from a server then i think this script has no more any use.
Just tell me how to copute it if i can do it in sh. What is the epoch ? you mean it is just the time of the cumby in millisecond from the 1 january 1970 as in Java (i develop using java at work). I don't think chumby has the correct time using internet synchro at the millisecond, it is never exact the time so the MD5 hash is approximative, no way to know wich hash is the youngest, you can only know that they are not the same.

tonigth i will speed up my script, it is still crap on the parsing of HTTP parameters, too many awk kills the awk ! Maybe i try it in perl. What kind of other scripting lenguages are available nativly ? or intalling a small exe or lib ?

I have put a 4Gb usb drive inside the chumby using a hub (i am in france so the warranty i dont care about).