GNOME Bugzilla – Bug 649092
inconsistency in the style of statistics fields
Last modified: 2011-05-03 00:13:59 UTC
In some statistics fields, the description columns or cells are in italic with colons (Principal Component Analysis) or without (Normality tests, histogram), and not in italic in others (descriptive statistics, ANOVA).
Do you have any opinion whether they should have a colon or not?
I prefer without, because * it would be more consistent with double-entry tables like in ANOVA, * it is maybe more logical in a spreadsheet : the structure of the table already makes the link clear, more if the description is in italic but that's only my humble opinion.
Thanks, I tended towards "without colon" too. And I agree that they should be in italics. The latter we can do any time, the former may have to wait until we branch since it affects all the translations. If you like to help, we need a listing of all the tools that use colons and a listing of all the tools that are not using italics.
Here is a list that should be complete. Uses colon : * Kaplan-Meier estimates tool (only 1) * Regression (only 1) * Principal Component Analysis * Sign tests (one ore two samples) * Wilcoxon Signed Rank Test (one ore two samples) * Wilcoxon-Mann-Whitney test * Contingency Table test (homogeneity and independance) Doesn't use italic at all : * Sampling tool (lacking for the title of the input) * T test * Z test * F test * Descriptive statistics * ANOVA Only partially uses italic : * Rank and percentile * Frequency table * Histogram (lacking for the title of the input) Some other oddities : * Exponential Smoothing Tool : display of alpha and gamma (should be on two cells ?) * Wilcoxon-Mann-Whitney Test : bold title * ANOVA Two factors : bold lines
Doesn't use italic at all : * Sampling tool (lacking for the title of the input) * T test * Z test * F test * Descriptive statistics * ANOVA These are fixed now.
Only partially uses italic : * Rank and percentile * Frequency table * Histogram (lacking for the title of the input) These are fixed now.
* ANOVA Two factors : bold lines Multiple factor ANOVA tables virtually always have these lines. Would you suggest to have them thinner?
Some other oddities : * Exponential Smoothing Tool : display of alpha and gamma (should be on two cells ?) This is intentional since the cells are usually very wide. I did fix some italic issues in the case that the data had no labels. * Wilcoxon-Mann-Whitney Test : bold title changed to just italic label for consistency with all the other tools. * ANOVA Two factors : bold lines changed the lines to "thin"
Uses colon : * Sign tests (one ore two samples) * Wilcoxon Signed Rank Test (one ore two samples) * Wilcoxon-Mann-Whitney test This has been fixed (in the Gnumeric source and the keys of the translation files). Note that the translations may still contain the colons. So if you are using a locale other than C this fix may take a long while to become visible.
Uses colon : * Kaplan-Meier estimates tool (only 1) * Regression (only 1) * Principal Component Analysis * Contingency Table test (homogeneity and independance) This has been fixed (in the Gnumeric source and the keys of the translation files). Note that the translations may still contain the colons. So if you are using a locale other than C this fix may take a long while to become visible. This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report.