[Digikam-users] strange behaviour of save and album in digikam 2.5

Marcel Wiesweg marcel.wiesweg at gmx.de
Sat Mar 31 18:20:28 BST 2012


> 
> Thanks for the explanation. I wonder how many would share your wish to
> save a copy without needing, or in fact being able, to specify the
> filename for the copy? For me, for anything other than Save, whether it
> is versioned or not, I would expect and want to see a dialogue box to
> specify name and location of the new file.

This sounds like you want to switch off non-destructive. You can do this.

In the purest form of non-destructive editing, only the commands are saved, no 
new pixel data, and the commands are reapplied when rendering the file. There 
are examples in Google and Nikon software. We did not want to take this road, 
applying complex operations on the fly does not seem feasible. So we chose a 
road where we still have actual files on disk - the original, the tips of the 
branches, and, depending on configuration, intermediate versions. Still, this 
is not old save+save as. There is no need for me to specify a filename or a 
location, I do not want to care. Digikam knows the versioning tree and can 
show me the original and the edited versions.

> 
> > So I thought I needed a different wording, still one short enough to fit
> > into the toolbar, and to emphasize that it's all slightly different with
> > versioning enabled. With versioning disabled, you're back to Save As
> 
> It is the very fact that versioning is enabled that is making the
> wording confusing, at least to me. IMO a new version of a file should
> still be logically attached to the original image whereas you are using
> it for creating a new file that is not logically attached to the
> original image. Perhaps 'Save changes as new image' would be better.

For us technically inclined the word "create a new branch" has a meaning which 
we know from svn/git and is what we do here. But for the average user, I fear 
this metaphor would lead to confusion.



More information about the Digikam-users mailing list