[GNC-dev] [6f3499590 Usability improvements for Bayes editor window] (was: gnucash maint: Multiple changes pushed

[Christian Stimming]

Zitat von Geert Janssens <geert.gnucash at kobaltwit.be>:

> Op vrijdag 18 januari 2019 23:54:07 CET schreef Christian Stimming:
>> commit 6f34995901dcfc999c675e5a4bc095eaf52a2d6f
>> Author: Christian Stimming <christian at cstimming.de>
>> Date:   Fri Jan 18 23:32:31 2019 +0100
>>
>>     Usability improvements for Bayes editor window
>>
>>     Remove the "Are you sure" question as it is simply annoying but does
>>     not help.
>
> Christian,
>
> In what way does it not help or is it annoying ?

Dear Geert, in this particular case it just doesn't make any sense to  
add another are-you-sure question before the actual deletion.

First, the to-be-deleted data itself is not any user-entered data, but  
only indirectly the user decision on previous imports that will guide  
suggestions in future imports, but no actual decisions right now and  
only suggestions for future imports. In other words, there is only  
very little to use here.

Secondly, there is no ambiguity in the potential buttons that a user  
could press. Contrast this to the register's "delete" buttons, where  
either a full txn or a split of a txn could be deleted, and the  
"delete" buttons in the register are one among a long toolbar list.  
Instead, here in the imap-editor-window, there is only exactly one  
button "delete", and it deleted exactly the selected line.  Also, the  
are-you-sure question does not add any information at all. In the  
register, the question highlights the fact if there are already  
reconciled splits and such, but in this case, the question is only   
the kind of "you pressed this button; did you really want to press  
this button?"

>
> I don't use bayes data so I don't know the specifics of how this works.
> However in general when gnucash is about to delete data we ask the user
> whether that's what s/he really wants to do. In many cases we also  
> provide the
> option to remember the answer.
>
> Is there an undo feature (other than restoring from backup) if the user
> deleted these bayes entries by accident ?

I did use this feature a little bit more. Sure, in an ideal world,  
there would be an undo button, but here and as long as we are in C  
this is not achievable. So we make our own assessment on how painful  
an accidental deletion of the wrong data would be. In this bayes  
editor window, as I said above, the data isn't any directly  
user-entered data, and its accidental deletion also doesn't change any  
user-entered data directly. It would add a little bit of extra clicks  
upon the next import, but then the data is there again. Because of  
this I decided this question is not needed here.

Regards,
Christian