On Oct 2, 2009, at 1:12 PM, Eric J Korpela wrote: > On Mon, Sep 28, 2009 at 11:15 PM, Raistmer <[email protected]> wrote: > >> >> It's a good point. Maybe Eric could answer if such checks were >> performed in >> lab before releasing SaH application. >> But anyway, such testing should be done only once per algorithm >> change, it >> should be done in project's lab and not on participants PCs. > > Sorry, I'm way behind on emails on this list. That is what is done. > We test each application release against a reference data file, and > against previous versions on random data files to validate the > application version. There are also continuous signals present in the > data at certain frequencies, so we can check that in aggregate the > results are valid.
Ah, I did not know that ... I have been asking this question for years and no one could answer it ... and in the test project which I monitored for longer than I participated always seemed to use only one test task for all local tests ... Regardless, alpha and beta tests don't catch all errors ... _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
