On 9/26/07, Gregor Hartmann <[EMAIL PROTECTED]> wrote: > I finally came up with an algorithm to produce 6 character IDs which can > be generated by a small perl tool. (which I will write soon) > > Its based on a CRC32 checksum which happens to be unique for todays sdf > files. (even including old strings which are still in the SUN database) > > Packing it with digits, letters and some special characters allows to > squeeze it in 6 characters. > > What will have to get decided is whether to use small and capital > letters and which special characters to use. > > My suggestion is to use small characters only since that's easier to > speak, say on the phone.
+1 > We need 41 symbols (digits) to squeeze it into 6 characters, so there > are five left after 10 digits and 26 characters. I suggest to use these > + - < = > > > I would love to also kick out the l and the o since they can be mixed up > with 1 and 0 but the we would need some more special characters which > IMHO is ugly. Well... since we won't use monospaced font on the interface, I don't believe it will confuse someone. And... I and O are easier to be confused with 1 and 0 than i and o, even on monospaced fonts... so it's better to leave the i and the o alone, than using other symbols. > So if no one objects I will use 10 digits, small letters and the above > mentioned characters to pack the IDs. > > what do you think? That's great! What we want most is to be able to build key-id builds to aid us in the translation process. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]