Bug 161934

Summary: "Top 10" item in Autofilter filter drop-down is ambiguous
Product: LibreOffice Reporter: Eyal Rozenberg <eyalroz1>
Component: CalcAssignee: Not Assigned <libreoffice-bugs>
Status: RESOLVED NOTABUG    
Severity: minor CC: heiko.tietze, telesto
Priority: medium    
Version: Inherited From OOo   
Hardware: All   
OS: All   
See Also: https://bugs.documentfoundation.org/show_bug.cgi?id=79140
Whiteboard:
Crash report or crash signature: Regression By:
Bug Depends on:    
Bug Blocks: 103512    

Description Eyal Rozenberg 2024-07-06 19:51:42 UTC 
When applying a column filter in an AutoFilter table, what does 

Filter By Condition > Top 10

mean?

Just reading that phrase, it could be: 

* "The 10 rows with the highest value"
* "All rows with any of the 10 highest values"
* "The 10 rows at the top of the table" ... ok, a bit of stretch, but still.

and I wonder how this is translated.

Anyway, as stragu also suggested in bug 79140, I suggest this be renamed into: "Top 10 Distinct Values" or "Highest 10 Values".

Note also that the phrase should "work" semantically for non-numeric values.
Comment 1 Heiko Tietze 2024-07-11 08:14:16 UTC 
Help [1] explains

"Displays the 10 rows of the cell range that contain the largest values in the cells of the current column. If these values are unique then no more than 10 rows will be visible, but if the values are not unique then it is possible for more than 10 rows to be shown."

The UI label is catchy and short and something like "Highest 10 Values" not an improvement, IMHO. => NAB

[1] https://help.libreoffice.org/24.2/en-US/text/scalc/01/12040100.html
Comment 2 Eyal Rozenberg 2024-07-11 18:09:01 UTC 
(In reply to Heiko Tietze from comment #1)
> Help [1] explains

Heiko...

You know that help text is never an excuse for lack of clarity in choice of UI labels.

> The UI label is catchy and short

It is catchy, and it is short, but the meaning must also be clear :-(
Comment 3 ady 2024-07-11 19:00:45 UTC 
(In reply to Eyal Rozenberg from comment #0)
> * "The 10 rows with the highest value"

I sincerely ask... How/why would that be the meaning? This is within the context of AutoFilter, right? When you select one (or several) items, is the expected behavior clear? Are the specific row numbers in such case relevant? Is there any reason for "Top 10..." to behave differently? If the item were to be "Top 3...", would you understand it to be the same (highest) value that was first found 3 times in the list, and then only show those 3 (first listed) rows with the same highest value?


> * "All rows with any of the 10 highest values"

Well, yes. Just by trying it once, users should be able to understand the result quite immediately, don't they?

> * "The 10 rows at the top of the table" ... ok, a bit of stretch, but still.

Context matters. This is in the context of AutoFilter. Showing a "podium" (of 10) refers to a rank, as with any competition.

If a user needs to show the first 10 rows, whichever the value they have, then the specific column in which the "Top 10..." filter is used is relevant. For instance, a column could be an index, and showing either the "Top 10" or the "Bottom 10" would be relevant for such filter on that column.

The AutoFilter on a specific column works according to the values in that column – there is the issue of the display format vs the cell's internal value, but that's a different unrelated thing.

In the context of the AutoFilter feature, I don't see the potential confusion, and even if there was (e.g. for newbies that don't have experience with this specific filter), one simple try/test should be enough to clarify its intention.
Comment 4 Eyal Rozenberg 2024-07-11 20:23:41 UTC 
(In reply to ady from comment #3)
> > * "The 10 rows with the highest value"
> 
> I sincerely ask... How/why would that be the meaning?

Because it says "Top 10". Is can be ten values, or ten rows.

> This is within the
> context of AutoFilter, right? When you select one (or several) items, is the
> expected behavior clear? Are the specific row numbers in such case relevant?

They aren't, in this interpretation. They're relevant according to the third possible interpretation.

> 
> > * "All rows with any of the 10 highest values"
> 
> Well, yes. Just by trying it once, users should be able to understand the
> result quite immediately, don't they?

No:

* Because it might be the case that the top 10 highest values have one row each - and then, the two most reasonable interpretations happen to coincide.
* Because next time they use this filtering they might not remember exactly how the confusion was cleared up last time.
Comment 5 ady 2024-07-11 21:54:02 UTC 
Other spreadsheet tools use the same "Top 10" label, even when there are additional options within that label (hence allowing different results).

There might be bugs and possible enhancements regarding AutoFilter. FWIW, changing the "Top 10" label is not one of those, IMO.
Comment 6 Heiko Tietze 2024-07-12 06:05:58 UTC 
(In reply to ady from comment #5)
> Other spreadsheet tools use the same "Top 10" label...
Good point. Changing the name makes the function not easier to understand but adds confusion for users. => NAB
Comment 7 Eyal Rozenberg 2024-07-12 10:56:08 UTC 
(In reply to Heiko Tietze from comment #6)
> Good point.

It's a poor point, because:

1. People don't interpret menu item labels to that level of specificity based on their memory of the interpretation of similar menu items in other office suites. That would be true for large mechanisms, such as PivotTables, but not at this level.

2. If something is mis-labeled in another app that doesn't mean that mis-labeling should be replicated.

However - since there are no other proponents of this change and since the original motivation was a comment by another user rather than myself, I won't contest the NAB.