At 08:02 PM 6/2/2005 -0400, Jerome Hartke wrote:
>Based on testing for our clients, I believe that you are pursuing the wrong
>quality indicators. "Too cumbersome and time-consuming" suggests that you
>wish an easy answer to a complex question. Such answers invariably lead to
>misguided efforts, a false sense of security, and unsatisfactory results.

May I suggest that we need to define the outcome we hope for from such
tests before examining procedures? For example, we might want to begin by
saying that we require a set of measures of quality of recording and of
accuracy of retrieval over time. If such measures can be defined, then a
methodology for their assessment is needed; we should then be ready to
charge the manufacturers with making and publishing those measurements.

Over all of this, we need to face the limitations of time and cost. They
are related in that the measurements need to be made available before the
medium has become obsolete. Thus, if one postulates a methodology for
measuring CD-R failure rate over time which requires a decade of test, the
discs in question can be expected to have vanished by the time their
performance is known.

On a trivial level, it is not even clear that anyone has enumerated the
failure modes of interest or a fitting measure of "failure", given that
bit-level errors are expected in DVD recording.

[log in to unmask]