Bug Report: Pie chart result is 100% blue when 0 unique files.

[wwcanoer]

Registered version 5.20.1 (the latest one).
Win10 x64 on Lenovo T430s i5
---------------------------------------------------
ISSUE #1: Bug

I ran a quick test to ensure that I was correctly interpreting the three colors of the pie chart and got a surprising result when there were no unique files in my test.

With 16 files: 4 copies of 4 different files, 0 unique files.
The resulting pie chart legend was correct:

Files with duplicates: 12 (red)
Duplicates only: 4 (green)
Remaining files: 0 (blue)

The pie chart should have no blue (0/16), be 1/4 green (4/16), and 3/4 red (12/16).
But it's 100% blue!

When I added 2 and then 4 unique files, then the result was as expected.

(red) Files with duplicates: 12
(green) Duplicates only: 4
(blue) Remaining files: 0 or 2

But when I added 16 unique files, I was SHOCKED! Thus bug #2:
-----------------------------------------------------
ISSUE #2: The legend order/color changes!

I call this a bug too because I was so scared that DC incorrectly recognized the number of duplicate and unique files!!! (It's been many years since I tested DC on known test directories.)

The color order stays consistent: red, green blue.
But the label order varies because it is sorted by the result number of files large to small.

With 32 files: 4 copies of 4 different files, plus 16 unique files
The resulting pie chart legend was surprising!

(red) Remaining files: 16
(green) Files with duplicates: 12
(blue) Duplicates only: 4

I was shocked because, expecting a consistent legend order, I looked at the bottom item, expecting to see 16 but it said 4! I was so scared that DC had a serious bug because I am just starting a big cleaning of my many TB of USB drives. I checked the other tabs and then went back and noticed that the order of the legend text had changed but not the colors.

I expect that this is a bug because It makes the most sense to me that since we are trying to remove duplicates, they are red at the top, followed by the number of source duplicates, and then the unique files which make sense is green since good and no action required. So, the order and colors in my Bug#1 test make sense to me. The changing order doesn't.

I think that best practice is for a chart legend to stay in a consistent color and order because a user will run many different tests and wants consistency so that they know at a glance what the chart means. If the color meaning keeps changing, then one has to stop and think about what it means.


(Can't upload images? If you want, I can email them but this should be easily replicated because it easily repeats on my installation.)

[wwcanoer]

I see now, you probably use the same code for the Duplicate Files pie chart as you do the File Type pie chart. For the File Type pie chart, it makes sense to sort the legend by the number of files (large to small) and the color probably doesn't matter much because people won't remember 12+ color assignments, although even there it would be better to keep consistent colors. (Doesn't matter for an occasional user but a heavy user might start remembering the colors if they are consistent.)

For the Duplicate Files pie chart I think that it make more sense for the order and colors to be consistent because you will be checking that result after every run to see the numbers change and I suspect most people would like to see a consistent order as they watch the top number drop as they clean duplicates. Could make sense to have the "remaining files" at the top because you want to see that number increase while the other two decrease. (unless you're aim is 2 copies of each file, but even then, I think that the below order and language makes the most sense)

(green) Single files
(blue) Duplicate source files
(red) Duplicate copies

Note: In the manual you don't show the Duplicate Files chart, so you don't explain the language of that legend.
https://www.digitalvolcano.co.uk/duplic ... ng_tab.htm?

[DigitalVolcano]

Yes, point taken. It does use a standard colour chart depending on the order - which as you say can isn't helpful in the duplicates pie chart.
Will log as needing a tweak in the next update.