Observation Oriented Modeling
Software Bugs/Glitches

OOM is new software programmed solely by James Grice. While we have tested the software thoroughly and used it to analyze well over 100 different contrived and genuine data sets, bugs are likely to remain. Hopefully, all of the major bugs have been detected, and only minor bugs will be reported here.

Version 2 Bugs

  1. "Bug" #2 below is not technically a bug. Each pattern must be printed individually.


Version 1 Bugs (Fixed in Version 2)

  1. The c-value is reported as "< 0.00".  This is not a bug. Increasing the decimal precision will show that, for instance, the c-value is less than .001 or some smaller number. The decimal precision can be changed by selecting  Options --> Decimal Precision --> Text Output  from the main menu. Please note that the analysis of interest must be run again before the results will be shown with the increased or decreased level of decimal precision. See the FAQ for more information about the c-value.

  2. Pattern Analysis (Crossed and Concatenated). Patterns can be defined in three dimensions. For instance, two orderings might be crossed to create a 2-dimensional matrix, and then Gender might be defined as the third ordering. Patterns can then be defined for the males and females separately in the Pattern Definition window. The pattern is thus a 3-dimensional pattern. The Print Pattern Image and Save Pattern Image options (buttons) do not function as might be expected. Specifically, only the visible layer of the pattern is printed or saved when either button is chosen. For instance, if the first 2-dimensional grid represents males and is visible, only the 2-dimensional grid for males will be printed or saved as an image file. To print or save the female 2-dimensional pattern, it must first be made visible and then the print or save pattern image buttons chosen. This is not a functionally important bug since the patterns are still defined and saved appropriately, but each layer must be printed or saved separately.

  3. In the Pattern Analysis - Concatenate Observations output, the c-value text is sometimes corrupted, as follows:

        Observed Proportion of Matches : 0.36
        Number of Randomized Trials : 1000.00
        Minimum Random Proportion of Matches : 0.17
        Maximum Random Proportion of Matches : 0.29
        Values >= Observed Proportion : 0.00
        Matching c-va<ue : 0.00

    Notice how the last line is corrupted. It should read "Matching c-value <: 0.00". With 1000 trials, this means the c-value is less than .001 (see #1 above).

  4. In the Pattern Analysis - Concatenate Observations option the ordered observations to be concatenated must be selected prior to selecting the  "separate by" ordering. An error message appears when the "separate by" ordering is selected first, followed by the concatenated observations.

  5. The [Stop] iterations buttons are not functional in the Pattern Analysis Crossed and Concatenated options. Consequently, once the [OK] button is selected and the randomization test begins, there is no way to interrupt the analysis except by terminating OOM through the Windows Task manager.