[Digikam-devel] Fwd: [Digikam-users] Metadata and IPTC support

[Joern Ahrens]

We should keep Caseys mail in mind, when we start developing the next version.
It is very interesting but sadly nobody of us found any time to answer....

----------  Forwarded Message  ----------

Subject: [Digikam-users] Metadata and IPTC support
Date: Wednesday 09 November 2005 05:59
From: Casey Finnerty <caseyfinnerty at rogers.com>
To: digikam-users at kde.org

To the Digikam development team and community:


I am in the process of migrating to Ubuntu/Kubuntu Linux and have been
testing and comparing Linux photo management software. As a user who very
much appreciates the community aspect of open source software, I would like
to share my thoughts on digikam re: photo annotation/metadata, in the hope
that these ideas may be useful for future development. Please excuse the
length of this email, but I thought I'd share my thoughts after studying
metadata storage options over the past year.

Based on what I do with photos, I have the following needs for a photo
manager (in addition to those already met by the great features offered by
digikam and other photo managers).

*****************
NEEDS:

1. Enter metadata once only!

2. "Near universal" standard for tags and comments,  so that this metadata
can be read by other applications and photo services.

3. Have metadata written directly to the image file (IPTC editing?), so that
it cannot be separated from the image.

4. Easy export to photo-hosting and printing
(Flickr/Smugmug/KodakGallery/etc.) services, with no need to re-enter tags
or comments for each photo. Admittedly, only Flickr, and possibly Smugmug,
read IPTC and EXIF keywords and comments. Digikam's Flickr export results in
the EXIF comments being entered at Flickr, but having to re-enter tags in
the export dialog is inadequate. For example, what if I'm exporting to
Flickr multiple images with different tags?

5. Fast searching of tags and comments. Digikam does a pretty good job of
this with its separate database file(s) and saved searches.

6. Powerful, flexible printing options. Print pictures with or w/o titles
and/or comments. Settings for print and paper quality. Digikam is ok for
printing, but there is room for improvement re: settings for print/paper
quality. Perhaps this is a wider KDE issue. (As a Linux newbie, I don't
know.)

7. Easy archiving of photos onto CD/DVD, with a searchable catalog of the
metadata for those photos archived on removable media, even if they are not
inserted in drive. Bonus for thumbnails returned with search results.

****************

At present, I feel that the best way to address the goals of archive
indexing and "universal" readability is to have metadata (e.g. Title, tags,
comments, etc.) stored in the IPTC and/or EXIF fields of each photo. To
support fast searching, this same metadata should be mirrored in Digikam's
database. Ideally, photo files could be scanned for IPTC data at startup,
and the new metadata entered into Digikams's database. Tagging photos and
entering comments could retain the existing approach, with a dialog box or
preference setting offering the option to write the metadata to the
appropriate IPTC fields. (I suggest that comments be written to the IPTC
caption field, and tags written to the IPTC keywords field.)

My current photo management is done on a Mac, using Graphic Converter for
photo renaming, date fixing and IPTC metadata editing. Graphic Converter has
a nice photo browser, and IPTC keywords can be entered by double-clicking
entries in a keyword list window. Keywords are written to each selected
photo with every click, though, which is somewhat slow (i.e. Spinning wheel
time). It would be better/faster if keywords could be quickly written to a
database file, then to the entire batch of individual photos with a click at
the end of the work session.

For archiving, I scan photos on CD or DVD with CDFinder, which can index
IPTC metadata (though it does not generate or save thumbnails in its
database).

On Linux, the only photo manager I have found that does complete IPTC
editing is Martin's Picture Viewer (Mapivi,
http://mapivi.sourceforge.net/mapivi.shtml). Mapivi will also index and save
to its database metadata on removable media. While this program is quite
feature rich, I also find it very slow and lacking some of the nice features
and interface found in Digikam. My preference would be that Digikam
incorporate the features I'm suggesting.

In terms of indexing archived photos, it would be very useful if digikam
could scan image files on CDs or DVDs and add the filenames, metadata,
location (and thumbnails?) into its database. Alternatively, if a CD/DVD
cataloger for Linux (e.g. Katalog) already does this, I'd like to know.

In closing, I'd like to address some of the concerns I have seen regarding
the use of IPTC for metadata entry/storage. After googling "digikam iptc", I
saw that this past summer the developers have been discussing and working on
an implementation of iptc support, though there was some debate over the
wisdom of writing metadata to IPTC fields. Most of the criticisms focus on
IPTC being limited to jpg and tiff, with no support for png, and the
uncertainty re: the persistence of the IPTC standard.

In terms of percentages, it is my impression that the vast majority of
consumer digital cameras utilize the jpeg format, with some offering tiff
support. I suspect that very few consumers use tiff for initial image
capture, due to limited memory card space. Most professional cameras (i.e. A
much smaller market share) utilize different types of RAW format, which I
understand varies among manufacturers. My conclusion is that at present the
jpeg (and tiff) formats are the most common, and therefore the IPTC
restriction to these formats is not a strong reason to reject it. For good
reason, many people convert their images to "lossless" png for photo
editing, but as long as the png version is saved in the same directory as
the original jpeg it should be easily located by searching the IPTC metadata
in the original file.

In terms of persistence, my understanding is that a new IPTC core standard,
compatible with XMP, has been developed. (See http://www.iptc.org/IPTC4XMP/
for more information.) I think we all agree that we'd like to "future-proof"
our metadata as much as possible, in order to avoid having to choose between
the pain of losing it or re-entering it for every photo in our collection.
Because IPTC has so many intrinsic benefits (many fields to hold metadata,
read by many applications and photo services) and a roadmap has been
developed for XMP compatibility/conversion, I think it is currently the best
choice for holding metadata for long-term accessibility.

I'd like to hear from the developers and other users their thoughts on these
ideas. If some of the ideas in this message are planned for implementation,
I'd love to hear about it too. I'm planning to build my first Linux box
soon, and I'd like to know when I can start moving my photo management over
to it!

Finally, let me thank the developers of digikam for putting so much hard
work into such an excellent package. Your efforts are sincerely appreciated
by a user who has a young daughter (and a rapidly growing photo collection
as a consequence). Keep up the great work!

Casey
--------------------------------

Casey M. Finnerty, Ph. D.
Guelph, ON
Canada