Turbo Pascal Error 200 bug on very fast (>=200MHz) computers - Turbo Pascal (which we use to compile our DOS programs) has a bug on fast computers: all programs generate a run time error 200 during start up when Turbo Pascal initializes the delay procedure. Patched CoHort programs can be downloaded for free (see the latest version). If you need to fix Pascal programs which you have written, see Frank Heckenbach's web page. (posted 8/97)
No "Year 2000" Bug - All versions of CoHort programs
(since 1.0) are Y2K compliant. CoHort programs store dates as
days-since-Dec. 30, 1899, so there is no problem storing any date,
including dates after 1999. CoHort programs have always been aware
that 2000 is a leap year and that Feb. 2000 therefore has 29 days,
so they can correctly do calculations
with any dates, including dates before and after Feb. 29, 2000.
Here's a test you can do to reassure yourself that this is true:
go to the data file editor in CoStat or CoPlot and enter
the following values (using the MM/DD/YYYY format for entering dates):
"12/31/1899" (the resulting value is 1, meaning 1 day since Dec. 30, 1899),
"12/31/1999" (the value for one century later is 36525),
"1/1/2000" (the value for 1 day later is 36526 --
it recognizes the year 2000!),
"2/1/2000" (the value for 31 days later is 36557),
"3/1/2000" (the value for 29 days later is 36586 -- it recognizes
that Feb. 2000 has 29 days!).
One warning related to entering dates: Since the first version, you can enter any date using the MM/DD/YYYY format (e.g., 12/31/1999). Thus CoStat and CoPlot have always been Y2K compliant if you use this format. But in versions before CoPlot 3.031, if you enter a date with the MM/DD/YY format (e.g., "12/31/99"), the program assumes you mean 19YY (e.g., 1999). Starting with CoPlot version 3.031 and CoStat version 5.031, for YY=0..49, 2000 is added (e.g., 12/31/15 is interpreted as 12/31/2015); and for YY=50..99, 1900 is added (e.g., 12/31/99 is interpreted as 12/31/1999). The latest version has this fix in it.
CoPlot NChannel Graph - Sometimes, when the number of windows should change, it doesn't (for instance, when you turn off one of the datasets). The data may be plotted correctly, but the Y axis Labels are in the wrong place and the old number of windows are shown. An easy solution: choose Graph Automatic changes Divisions Automatic and Labels Automatic.
CoStat Lattice ANOVA - For lattice designs, CoStat calculates the unadjusted treatment MS and F values when it should calculate the adjusted treatment MS and F values. Because the calculation is beyond the scope of the "language" used in CoStat's .aov files, there is no easy fix. For now, you must do the ANOVA by hand (see "Statistical Procedures for Agricultureal Research" by K.A. Gomez and A.A. Gomez). (posted 12/3/97)
Go to: CoHort Software | DOS Versions