[Shotwell] The old DB vs File-system and Destructive vs Non-destructive management of photos.

[Bruno Girin]

Hi Svetoslav,

Thanks for your input, this is very interesting! I've added some
comments below on some of your points.

On Sun, 2010-06-20 at 06:09 -0700, Svetoslav Trochev wrote:
> Hi Everyone,
> 
> Please forgive me if I am stepping out of line a bit and because I am
> new and don't know the background. For many years I am struggling to
> find a good tool to manage my photo work flow. I am still stuck using
> a multiple tools and some of them are close source and I hate them.
> They hold me hostage to their ecosystem without solving my problem. As
> result I spend way too much time to manage my photos instead of doing
> something fun like creating and editing them. I am happy that I
> discover the Shotwell right at the moment when discussion about photo
> management is going on. I would like to share my ideas and tap into
> community wisdom in order to solve the "my photo management" problem
> once and for all. So let start. My work flow is:
> 
> Phase I:
> 1. Get an image. (cameras, scanners, web and etc.)
> 2. Store them in some file-system structure.
> 3. Enter basic meta-data. ( Event, People, Purpose, Source, Copyright,
> ratings, and etc)
> 4. At this point this is my 'Digital Negatives' (DNs)
> 5. Remove useless crap.
> 6. Basic non-destructive editing usually using RAW converter ( Lens
> corrections: geometric distortion and chromatic aberration, tilt, and
> perspective, exposure, and color correction )
> 7. Archive (Network share and Removable media)
> 
> Phase II:
> Image processing. This is very different based on the job. It includes
> things like selecting the right images, HDR, 3D and etc. Using the
> right tool for the job. Most often it is destructive editing
> 
> Phase III:
> Sharing the images. Again it varies a lot
> 
> As you can see I need both DB and File-system management. Because I
> want to be able to remove the DNs from my local file system, but to
> continue to be able to search the archive and see preview. This only
> could be done by DB. At the same time I need to manage multiple
> 'prints' as part of Phase II. This only could be done in different
> project folders and should share meta-data with other tools. And
> finally to maintain multiple share options by creating galleries this
> is most well done by DB/Export. That way I don't have to maintain
> multiple backups of the same photo.

There are lots of interesting suggestions here. I think the concept of
having multiple prints could be very useful, even for users who are not
as advanced as you are. For example, any user would understand the
benefit of being able to have, say a "black & white" print and a sepia
one and compare between them.


> 
> When it comes to editing I need both again. Non-destructive edits
> organized in 'recipes' The recipes should be attached to the
> RAW/original file and I would like to copy recipe from one file and
> add it to one ore more files. This feature speeds my work multiple
> times. That is the main reason why I still use original software from
> Canon and stuck with Windows. In all other cases I really like how
> Shotwell manages the edits and the use of DB. Probably would be good
> idea to be able to export edit data if needed to other tools and
> managing the steps like version control between different tools when
> destructive editing is needed. So new feature would be needed I think.
> Other tools call it stacks where you see the last result, but you can
> go back and see previous steps.

In practice, what Shotwell does is create a recipe to re-create a
particular set of edit on top of an existing negative. Being able to
copy that recipe from one photograph to another to apply it in batch to
a collection of photograph would be very useful.

The concept of stacks could actually be extended to allow branching in
the stack, in order to support the concept of multiple prints (and
derivations thereof). In fact, you're right, the best way to see this
would probably be to model it on version control tools, with added
visual goodness.


> 
> I understand why proprietary tools never going to provide this unless
> it cost $$$$ and needs to be updated every year for $$$ more just to
> add new feature or to fix a single bug or even worst to continue to
> access your photos. I strongly believe that FOSS is the right way to
> fix this and I am bit surprised that we have not done this already.

I agree with you, FOSS is the right way to fix this and I believe
Shotwell is a good platform to do so. It certainly has the potential.
What you've said above has given me a few ideas. I need to clarify them
and put them down in writing. Once I've done that, I'll share it with
the list so that you can tell whether it makes sense.

Bruno