Tagging & Meta-Data

The current tagging hype seems to be everywhere. Mostly observed at flickr, technorati and del.icio.us, but basically it is all about meta-information, which will finally arrive in most day-to-day uses, be it through the beforementioned web-applications or apple’s cool new search feature for MacOSX Tiger (10.4). Not all meta-information is as valuable as it seems, but hey, we all need to start somewhere.

in DevonThink

I have been thinking how I might want to increase the value of my stack of literature I have here for my thesis and DevonThink was one of the first choices, as it “creates the interconnections” quite easily with its superb “see also” function. Its Auto-classification system does not work for me yet, but then I do not really have a good structure of the papers in my database yet and I was thinking along a tag based structure as well.

There is a discussion going on here about how to group single documents in multiple groups (as could be done with replicates). And I suppose there are two ways of doing such a thing:

1. Add keywords to my files via the file’s comments.
2. Create replicates in my “tag” groups.

Number 1 is too much work, i think, as I have no way of making this easy to do, Number 2 is just more intuitive and also might provide the classification features I am looking for. However: It would also be really cool if I could tell DevonThink a number of keywords that have aggravated meaning for DTs relationships. If I tell DT that Files A and B both are about “cooking lobsters” and there is not even a single mention of “lobster” in both files I still want DT to turn up both files when I search for it, as well as group other files togethere with them that do contain the term. Is this doable?

I realize there would be a possibility for tagging with keywords that worked through AppleScript - but I have not learned to think in AppleScript yet; I envision a simple interface for keyword usage as in iPhoto5. (in fact I would love to have this kind of feature in iTunes as well).

visualizing connections and relationships

What would also be cools is a feature hinted at by Christian Grunenberg here which would be really amazing:

If you’re using lots of replicants, then the structure of the database isn’t hierarchical anymore - you’re creating a net. Therefore the only issue might be that there are no other views (e.g. maps) available yet. But they’re in the pipeline

Which also makes me wonder: I have most of the literature I am working with in a bibtex file, associated with keywords. Is there any way of parsing that file and creating such a thing as a mindmap based on these keywords? Grouping similar papers (with like keywords) and possibly showing relationships between papers that I have not seen before (as there are too many papers to have them all in my head) and/or to just serve as a topic based list of articles in my database…? Somehow reminds me of what Lion is talking about in his book on “mapping his thoughts”.