I was just demonstrated a pretty hefty bug in OpenOffice’s spreadsheet component which a) completely violates the principle of predictability in user interfaces and b) makes this application absolutely unusable for many spreadsheet tasks.
Consider the following scenario: You have a large set of data points in a column and want to visually filter for all rows with a specific value and then overwrite these with a new value. Such filtering tasks are so common that there’s a dedicated command for them: the “AutoFilter”. To overwrite the filtered data, users then typically enter the target value in the first row they want to change and fill the succeeding rows by click-dragging the mouse from the first row downward.
As you’d expect, you have just updated all selected values in the filtered dataset. What you don’t expect: OpenOffice has also replaced all values in the invisible in-between rows which were intentionally hidden by the filter!
I’m sure you can immediately come up with a scenario where such unexpected application behavior can be disastrous in a business environment, especially since the outcome is not immediately visible in large data sets. What’s more, this defect has first been reported 8 years ago and is still not fixed!
Even though I am an advocate of Free and Open Source Software, I really think that the OpenOffice team needs to get their act together and better prioritize their tasks. Until this issue has been fixed I will no longer recommend OpenOffice to businesses.Update 2010-11-10: Apparently this bug is finally fixed in OpenOffice 3.3. Phew, it only took 8½ years.