Digital archiving - part 2

Part 1 in this little series of posts is here.
Digital Asset Management has emerged as a separate field of expertise. One quite authoritative source for describing the issues seems to be The DAM book by Peter Krogh. I have not read the book myself, but from looking at the index on the website I think it's likely that I will. Soon.
Meanwhile, I guess I must risk emphasising the trivial and ignoring the salient, by making my own list of things I find essential for a good system.

1. Magic words
I know the last time I made a real effort of putting keywords on my images was in August 2007; eighteen months ago. I hate it, and even posted about it here. So I need a system that makes this job as simple and rational as possible.
In particular, I would be very partial to a system that allows setting whole sets of keywords at once. If I, for example, add the keyword "Eagle" to 200 shots, it should automatically add some associated keywords like "bird" and "predator"; provided that I have made "Eagle" a predefined sub-category of both.

2. Version handling and bundling
I shoot RAW, publish JPG on the web, and print from TIFF. Then maybe I want JPGs in different resolutions for other purposes. All of which are versions of the same image. I want my archive to group and deal with all versions of one image together, and to display only ONE thumbnail for the whole heap. Likewise, I would like it to bundle related images together from eg. rapidly recorded sequences as in sports, or tiles of a panorama. It should also have a clever way of letting new versions or whole bundles inherit metadata (eg. keywords) from one chosen master-file (eg. the assembled pano, the best of a burst series, the raw file when developed, etc.)
3. Backup, restore, and migration
The importance of having a system that will survive such major changes in hardware or software cannot be underestimated, in my opinion.
It should be easy to create a backup, and it should be easy to restore from that backup.
It should also be realtively simple to migrate to another archiving system without losing the metadata.

4. Support for offline archives
"Offline archive" means different things in different contexts. I am thinking of two different meanings:
4a. Disconnected storage media
The software should retain a thumbnail and the metadata for images that are stored on disconnected media such as DVDs, USB-drives, NAS-units, and so on.
4b. "Branch" archive on a laptop
The software should be able to upload parts of the archive to a laptop. For example a prepared slide show. It should also be possible to import from the branch archive back into the main archive in a simple way. If modifications of existing files are discovered, the version handling (see pt. 1) should deal with it.

5. I could go on
to add a lot of other things too. However I feel that maintaining metadata, handling versions, and versatile backup are the most important functions of an archive. Other things, such as intuitiveness of the interface and overall convenience will emerge in overall evaluation of a product anyway.

In the next post I will short-list some candidates that seem to have a lot of potential.