Information Navigation

When you want to find something, there are a few ways. The current systems allow us to group information into files, which are put into folders. To find something, we navigate a tree-structure until we reach the data nodes - files.

I think an alternative system would be tags, metainformation, and inherited tags and searching, as opposed as filenames and browsing. Not that file names are not allowed, but any file name can simply be another tag. What differentiates a tag from a file name?

• you do not have to know all the tags to access some piece of information
• you can emulate a file structure through inherited tags. that is, when you place some information into one tag, it automatically recieves other tag attributes
• metainformation may be extracted from the file and searched. when files are written, they will be run through metainformation strippers, that extract useful information for searching. for example, text documents may have special metainformation extractors that extract titles, important words, key words etc.

On the other hand, tags does pose a computationally more intensive way to find files. Data structures such as hashes or other relational database techniques might have to be used to implement this sort of file system.

To access information, I envision that we would use both selection and refinement. We can give a set of search terms to the information navigator. If the search is enough, then we will only be faced with one file, or possibly a group of homogeneous files (with similar structures). We can view these files because any type of information can be viewed. We can then either directly (via a function explorer) or indirectly (through a user-interface that you can load) call some functions to bear on the information.

Interestingly, a function explorer is not any more direct than a user interface - a function explorer is simply a viewer that views a list of all functions grouped into hierarchies or otherwise.

An interesting failed project