Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 312641

Summary: Change label and icon informing about capped comparison results
Product: [Eclipse Project] Platform Reporter: Pawel Pogorzelski <pawel.pogorzelski1>
Component: CompareAssignee: Pawel Pogorzelski <pawel.pogorzelski1>
Status: VERIFIED FIXED QA Contact:
Severity: normal    
Priority: P3 CC: daniel_megert, gro.espilce, markus.kell.r, susan, Szymon.Brandys, tomasz.zarna
Version: 3.6Flags: markus.kell.r: review+
daniel_megert: review+
susan: review+
Target Milestone: 3.6 RC1   
Hardware: PC   
OS: Windows XP   
Whiteboard:
Attachments:
Description Flags
Patch_v01
none
Info icon
none
Patch_v02
none
Patch_v03 none

Description Pawel Pogorzelski CLA 2010-05-12 11:03:36 EDT
The message was introduced on bug 292831. The new message should be "Faster comparison algorithm used" as Markus suggested and a warning icon should be changed to info.
Comment 1 Pawel Pogorzelski CLA 2010-05-12 11:06:55 EDT
Created attachment 168163 [details]
Patch_v01
Comment 2 Pawel Pogorzelski CLA 2010-05-12 11:07:38 EDT
Created attachment 168164 [details]
Info icon
Comment 3 Pawel Pogorzelski CLA 2010-05-12 11:09:39 EDT
Markus, please review it. The effect is visible in attachment 168160 [details].
Comment 4 Markus Keller CLA 2010-05-12 11:15:06 EDT
Perfect, thanks! +1 for RC1.
Comment 5 Pawel Pogorzelski CLA 2010-05-12 11:20:01 EDT
Changes in HEAD. Marking as FIXED. Thank you for the review.
Comment 6 Dani Megert CLA 2010-05-12 11:25:02 EDT
Icon change is OK for me but changing the label is not OK: it destroys the whole idea of letting the user know that the result might not be accurate: no one would hover over an icon that tells you that a faster algorithm is used.

Please pull out the text change. Markus agrees on that.
Comment 7 Susan McCourt CLA 2010-05-12 11:45:29 EDT
I still find "Matching might not be optimal" to be pretty unhelpful/useless.

What about my text suggestions from bug 292831 comment 68?
Short message:
"False differences are possible"

Hover:
"To avoid long computation time a faster comparison algorithm has been used. As
a result, differences might be shown in the viewer where there are none."
Comment 8 Markus Keller CLA 2010-05-12 12:03:58 EDT
(In reply to comment #7)
+1, even better.
Comment 9 Dani Megert CLA 2010-05-12 12:07:22 EDT
+1
Comment 10 Pawel Pogorzelski CLA 2010-05-12 13:01:07 EDT
Created attachment 168201 [details]
Patch_v02

(In reply to comment #7)
> Short message:
> "False differences are possible"
> 
> Hover:
> "To avoid long computation time a faster comparison algorithm has been used. As
> a result, differences might be shown in the viewer where there are none."

I had a discussion with Susan about this. "False differences are possible" is not exactly a precise label in this case. When faster algorithm is used there can be no differences in places where Compare doesn't show any. The problem is that a more optimal matching can find diffs in totally different places (for example at the end of the file where faster compare finds them at the beginning). Those are both correct matchings but we take the one resulting in the smallest edit distance as the best.

We came up with following:

Label: Differences shown not be optimal

Hoover: To avoid long computation time a faster comparison algorithm has been used. As a result, the differences highlighted in the viewer may be larger than necessary.

Patch making the necessary change.
Comment 11 Markus Keller CLA 2010-05-12 13:10:45 EDT
> Label: Differences shown not be optimal

That does not sound like correct English. Shouldn't it be:

Differences shown may not be optimal
                  ^^^
Comment 12 Dani Megert CLA 2010-05-12 13:12:06 EDT
Or: The indicated differences might not be optimal.
Comment 13 Pawel Pogorzelski CLA 2010-05-12 13:15:09 EDT
Created attachment 168204 [details]
Patch_v03
Comment 14 Susan McCourt CLA 2010-05-12 16:43:28 EDT
in general I prefer 

"may not be optimal"

over

"might not be optimal"

but i think this is subjective on my part, I'm not sure one is more proper than the other.  I'm okay with the patch.
Comment 15 Pawel Pogorzelski CLA 2010-05-13 05:11:36 EDT
Patch_v03 in HEAD. Marking as FIXED.
Comment 16 Markus Keller CLA 2010-05-17 04:40:11 EDT
Verified in I20100516-0800.
Comment 17 Daniel Sokolowski CLA 2014-10-08 10:55:04 EDT
The info message implied to me there is a slow detailed comparison algorithm, and so for the past 30 minutes resulted me trying to find how to activate it - does one exist and how to enable it?

Thank you :)
Comment 18 Dani Megert CLA 2014-10-08 11:08:32 EDT
(In reply to Daniel Sokolowski from comment #17)
> The info message implied to me there is a slow detailed comparison
> algorithm, and so for the past 30 minutes resulted me trying to find how to
> activate it - does one exist and how to enable it?

No, there's no preference to ask for the non-capped algorithm.
Comment 19 Markus Keller CLA 2014-10-08 11:45:23 EDT
Yes there is a preference, but it's a bit hard to find:

Preferences > General > Compare/Patch > Tab: Text Compare > Disable capping when comparing large documents
Comment 20 Dani Megert CLA 2014-10-09 04:12:42 EDT
(In reply to Markus Keller from comment #19)
> Yes there is a preference, but it's a bit hard to find:

Indeed. I added a keyword now, so that it's a bit easier to find.
Comment 21 Daniel Sokolowski CLA 2014-10-09 22:40:43 EDT
@Dani - just curious what keyword and is it used in the search preferences dialog box?
Comment 22 Dani Megert CLA 2014-10-10 05:20:41 EDT
(In reply to Daniel Sokolowski from comment #21)
> @Dani - just curious what keyword and is it used in the search preferences
> dialog box?

Yes, while typing "capping" it will find the page, i.e. it will also find it when you enter "cap".
Comment 23 Markus Keller CLA 2014-10-16 09:34:39 EDT
Filed bug 447565 with a patch that adds a context menu to the capping message.