Alec Flett wrote: > I know this sounds redundant, but I think it would actually be helpful > to put "s" next to measurements that are in seconds, and % next to > percentages.. I mean actually in the table, not just in the header... > Also, adding a "+" to indicate an increase would help readability.
A screen real estate issue, but we might be able to fit that in. > Finally, it would be nice if we could reduce the number of digits, I > forget what its called in scientific notation, but basically that no > more than 3 digits of information would be shown, so that you could have > "192s" or "44.2%" or "1.04s" We actually do that on http://wiki.osafoundation.org/bin/view/Projects/PerformanceProject. See http://en.wikipedia.org/wiki/Significant_figures and http://en.wikipedia.org/wiki/Significance_arithmetic for more information. I am not sure if there are Python libraries that would format numbers this way (have to search). We've just opted for a format with three decimals because some tests pretty much require it to be able to see differences (doesn't make sense for import or startup tests). > I think there isn't necessarily value in knowing that the "median" > column is the median, or the average, or what - its just the current > measure... how about just "current" since really what we want to be > doing is distinguishing it from the 'm5' column? Yeah, we are working on being able to run multiple tests in single tbox cycle so that we can get quick feedback on how your change affected things, so what the current columns are may change. -- Heikki Toivonen
signature.asc
Description: OpenPGP digital signature
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Open Source Applications Foundation "Dev" mailing list http://lists.osafoundation.org/mailman/listinfo/dev
