What to do about ever-increasing sample frequency?

[haydnw]

Sample frequency increases by 0.01 seconds per sample, so at high frequency it will start to deviate after running for some time. Professional-grade scientific gear skips a sample or just resets the clock every 24-hours - do we just mimic that behaviour?

[lab4d]

hi, i was wondering this too, i have to add two microphones to it to track sounds passing by, and i was wondering how to get the best frequency out o those at least, perhaps if we could have two processes running, one for "normal" sampling, for sensor that really dont need high frequency, such as pressure, but also have a a very fast one for the mic, LDR and others.. Is that feasible? do you think its worth it ?
Perhaps we could only separate the publishing, meaning for example publishing on xively could have a defined frequency, but publishing on local http server could be instant, so we can do requests to it at a faster frequency..

Luka