Re: Adding "Find Duplicates" feature to F-Spot


Fisrt of all, thanks for the feedback :)

El dom, 19-06-2005 a las 13:04 -0400, Benjamin Kahn escribió:
> Hi Alvaro,
> 	This seems like a somewhat strange use case.  

Yes, thinking about it, the tag duplicate is something I have reached in
the first time because how it is implemented (as far as I understand in
the few hours I have seen the code) F-Spot photo list view. But I am not
sure that it isn't a correct path using this approach.

> I believe there are two
> times when I want to find duplicates:
>      1. When importing a bunch of new pictures, I want to locate the
>         duplicates and not import them.  (Duplicates in this case are
>         usually photos that I have ALREADY imported, not duplicates in
>         the collection I'm importing.  But it is possible that the
>         collection I'm importing as duplicates as well.)

Interesting. I am thinking in the second case, when you import a
collection with duplicates. But we need to cover also the already
imported case. Also, the user could want to have duplicates so we can't
block the possibility of having duplicates in a collection (we need a
user case to show this point ...)

So you import a new collection (with duplicates in the collection or
with duplicates with already imported photos). You need a way to find
the duplicates in the two use casess.

The user can go to the "Find->Find Duplicates" menu entry and she
expect, I think, that in some way a photo view appears with all the
duplicates. I feel that using the main photo list view could be a good
solution for that. So the photo list view in this case is filtered to
show only the duplicates as for example, when you filter it to show a
date range. The user then selects the duplicates she want to delete and
commmits the removal process. 

In this scenario what we need then is the ability to show a list of
photos, a query which does it. 

>      2. Finding and immediately acting on duplicates in the photo
>         collection already imported.  I usually only want to do this
>         because I forgot to weed out duplicates when importing.  (Maybe
>         the feature wasn't implemented yet?  Maybe I wanted the new
>         photos I'd imported, but not the old ones?  Who knows.)  Here, I
>         want to search for duplicates and right away be looking at all
>         of them so I can delete them.
> I'm not sure why I'd ever want them to be tagged with a special tag.  I
> guess that might be something I might do AFTER doing the duplicate
> search and seeing only the duplicates.

Yes, the actual idea I have in mind (as I said maybe thinking in how
things are implemented right now, and not in the *best* way to do it)

1. Initially there isn't a Duplicate tag at all, so users that don't
need this feature don't suffer in the UI for it.

2. When the user select "Find->Find Duplicates" the:

	2.1 A new tag Duplicate is created
	2.2 All the duplicates photos from the currect selection or all the
albumns if no selection is done are found and marked with the duplicate
	2.3 The Duplicate tag is marked and only the duplicate photos are
shown, as you can see in the screenshot:

	2.4 The user then can travel thought the duplicates, delete some of
them, doesn't touch other ... I think it is a powerfull model to manage
duplicates, but maybe, overkill for the normal user.

I have some problem with this model, that I feel works in a very natural
way for how F-Spot works:

1. The duplicate tag is of no use when the user clean all the
duplicates. She can remove it easily but you know, normally, the
Duplicate tag will be in the tag list in the 90% of users forever once
they have used the Duplicate feature.

2. If you mark the original and the copy (you don't know which if two
photos is the original ... maybe looking the date but ...) all the
photos that have a duplicate in the past will be marked with the
duplicate tag. The user can remove the tag later but we can't put in
user shoulders something we have to solve now with a nice design. If we
can find a way to remove the duplicate tag in the right moment, all the
duplicate tags in all the photos will be cleaned.

I have only looked in how gThumb does it and I don't like it because it
only works to find duplicates in folders in the disk. I haven't looked
at how Picasa2, iPhoto or Photoshop Album until I have tried to explore
some paths without looking at them.


-- Alvaro

> On Sat, 2005-06-18 at 00:28 +0200, Alvaro del Castillo wrote:
> > Hi!
> > 
> > In order to make easy to talk about the proposal I have created a
> > screenshot proposal so you can see how it will work.
> > 
> >
> > 
> > Cheers
> > 
> > El vie, 17-06-2005 a las 20:21 +0200, Alvaro del Castillo escribió:
> > > Hi guys!
> > > 
> > > This is the first time I write to the list so I try to be as polite as
> > > possible in the first steps ;-) (later also, of course).
> > > 
> > > I have a big album with about 8.000 photos and I plan to export all of
> > > them using F-Spot to my gallery [1] website. F-Spot has imported all the
> > > photos without problems but I have found in the import process some
> > > problems:
> > > 
> > > 1. I have lots of duplicates in my album and I will need to clean them.
> > > 
> > > 2. The export feature from F-Spot to gallery doesn't seems to be rock
> > > solid:
> > > 
> > >
> > > 
> > > So it is time to try to help. I have seen that the TODO list is very
> > > large, so F-Spot needs a lot of work, but working in 1. and 2. could be
> > > the best way to see if I find enought time and motivation to work in
> > > this great application.
> > > 
> > > So I am working right now in "1. Finding Duplicates".
> > > 
> > > The process could be very simple:
> > > 
> > > - In the Find menu add a new entry "Find Duplicates".
> > > 
> > > - When the new menu entry is selected I travel trought all the selected
> > > images, or all the images is there isn't a selection, and doing a MD5sum
> > > of the files contents, it is easy to see if two images are the same.
> > > 
> > > - I plan to present the user all the duplicate images with the original.
> > > All the duplicates will be selected so it is very easy for the user to
> > > remove them if she wants.
> > > 
> > > Trying to show the group of duplicates images only I have seen that it
> > > is easy to show images that are tagged, or it is easy to use a time
> > > range to show some specific images, but it is not as easy in the first
> > > travel trought to the code to show a specific list of images.
> > > 
> > > Thinking about the best way to do it maybe creating a new tag
> > > "Duplicate" could be a great and direct way to implent the feature. The
> > > user that wants to find the duplicates images just select the
> > > "Duplicate" tag. I will use also the "Find" menu entry because it is
> > > very natural for the user to go there to find the duplicates.
> > > 
> > > I plan to find sometime this weekend to finish the implementation so any
> > > ideas about the implementation proposal will be very welcomed.
> > > 
> > > Thanks guys for the great product you have created!
> > > 
> > > Cheers
> > > 
> > > -- Alvaro
> > > 
> > > [1] (sorry, it is in spanish
> > > language)
> > > 
> > > 
> > >  
> > > 
> > > _______________________________________________
> > > F-spot-list mailing list
> > > F-spot-list gnome org
> > >
> > 
> > 
> > _______________________________________________
> > F-spot-list mailing list
> > F-spot-list gnome org
> >
> > 

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]