Check Harddisk (c)1996 Marnix Timmermans

These are the changes that were made to Check Harddisk:

0.81    first beta. Only command-line interface.

0.94    first public beta with new and improved user interface.

0.95    - removed problems with harddisks on secondary controller
          (mismatch INT 13 info and IDE info);
        - added 'popup' help screens after error and finished testing to
          encourage people to submit their results;
        - added time-out in sectors per track if it doesn't work;

0.96    - fixed 'missing sector' problem;
        - improved stability and possibilities for fixing errors;
        - removed problem with sorting and loading results;
        - improved handling of timing deficiencies in full-testing;
        - faster detection of disk caching programs;
        - disabled testing in multitasking operatings systems entirely;
        - track estimate factor is 30 now instead of 25;
        - fixed truncating of model name;

        - fixed bug created caused by 'solving' truncating model name bug;

0.97    - fixed serious bug in testing which probably caused many
          crashes. (the avgspt was used in one of the calculations, but
          not yet calculated)

0.98    - changed dbf-structure. It's incompatible with the old format
          now...
        - major revision of trigger value and head/platters
          testing-routines;
        - added print-function in help-window;
        - added printing/exporting of summary list in browse window;
        - included better IDE-identify code;
        - included seperate Buf2Host testing program;
        - improved random seek testing: statistically correct and rounding
          to half milliseconds;
          In most cases this means that the random seek is now faster
          than in previous versions;
        - added 'simple' testing. If normal testing doesn't work it will
          switch to simple testing. Simple testing is somewhat less
          accurate and extensive, but still very useful because the most
          important numbers (sustained throughput, random seek, buffer
          to host throughput) are still there.
        - probably some other things I just can't remember;
0.99    - changed dbf-structure again (meaning that it's incompatible with
          previous formats)
        - completely rearranged testing
        - added performance indeces (max and avg)
        - rounded off sustained throughput to nearest 0.05 MB/s
        - rounded RPM to logical values
        - used new CPU detection routines
