[Digikam-devel] [Bug 114465] Wish: simpler entry of tags

[Jeff Lehrkamp]

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
         
http://bugs.kde.org/show_bug.cgi?id=114465         




------- Additional Comments From lehrkampj bellsouth net  2008-02-29 05:23 -------
Tags seem to be one of the biggest issues I have had with digikam.  I have used picasa in the past and the tagging of images in digikam seems extremely difficult.  

The autocomplete suggestion above would be a great improvement.

The remainder of this comment is as much question as comment.

If I understand correctly from reading the IIM and XMP specifications and what information I have been able to deduce online, the tags are saved in the metadata as single "keywords" in the IMM which will translate to the dc:subject in the XMP, and the parent-child relationship of the tags is internal to the database in digikam and is not stored in the metadata? (question).

If I am incorrect in the last assumption let me know and the remainder of this comment is moot.

I have been using picasa in the past and picasa has no such parent-child relationship with the tags.  A single tag text or keyword is "unique".  The tags can be assigned to any file at any time with no relationship.  This in itself can be a major issue as the number of tags increase, up to literally thousands of keywords.  The one saving grace of picasa was already discussed here, it has the capability of allowing a typed field to assign the tags, and also to filter by tag.  This still creates a problem with different tags being assigned if the text is entered differently, for example issues with plural words and words that start with the same first few letters.

I work with 20K+ files.  With multiple tags on each file, the tags are impossible to manage.

I was hoping that the parent-child set up of the tags panels would assist me with this issue.  

I'm not sure I can see where the flexibility to set up the parent-child relationships in the tags is really useful other than as a means to organize and manage the tags. In most cases the tags I set up are a descriptive piece of information in itself.  For example I have multiple albums and a single tag, say something like "new york", is descriptive in itself.  I may have a album of photos containing photos of an event that takes place in new york, and another album with scenic photos of new york.  New york as a tag is descriptive in itself, and with the combination of tags for filtering I can get to the files that I need efficiently.

With the current database set up, this requires me to either set up a unique parent-child relationship for these photos with new york being a child in each, or completely ignoring the parent-child relationship completely and just have a huge list of individual tags, like picasa, that are impossible to manage.

One of the issues I ran into was I "moved" one of my photo collections on disk outside of digikam.  When digikam reloaded the collection, any of the keywords that did not have a complete parent-child relationship (which, ooops, as I was used to using picasa, many did not) created duplicate keywords, outside of the parent-child relationships I had already set up, so I have duplicate tags (keywords) everywhere. 

Allowing the tags to be unique, is one possible solution I might suggest.  

If there is a use for the parent-child relationships that I am missing, possibly just allowing a flag, internally within the program settings and/or tied to the keywords in database, to flag the tags as being unique or not.  

Allowing tags to be unique, when files are imported into digicam, it would find existing tags within the database (similar to what picasa does) and just link the metadata tag to the existing tag in the database.  this would allow the tags to be grouped together by "dummy" tags as parents.  You could to set up parents something like "locations", "events", etc., then all the relevent tags could be grouped as children under those tags.  As the child would be unique, the parent tags would not have to be assigned as tags in the metadata.  

Currently the only way to group these types of tags together requires that all the parents of each tag also be tagged in the metatdata of the file.  Additionally If importing a file into digikam that already has existing tags, but may not match the current parent-child schema that I had set up, it creates duplicate tags, making the tags even more unmanageable.