How Many Hat Racks?

I've been reading Universal Principles of Design, and came across the "Five Hat Racks" principle, which states that:

"There are a limited number of organizational strategies, regardless of the specific application: category, time, location, alphabet, and continuum."
This system strikes me as being simultaneously over-specified and under-examined.

Alphabet, Continuum, and Time are all variants of the same theme: ordinal arrangement. These three are important, though far from exhaustive, qualities which a data set might be ordered by. In the book's example utilizing buildings, the Time variable is the date of construction, though it might just as easily be the time which sunlight hits the top floor of each in the morning. Likewise, Continuum uses the height of the building, though number of steps or number of windows could also easily be ordered.

Category is a second low-level method, although I think of it as 'grouping'. Grouping is inherently unordered, although in many cases the data points within a group are ordered at the next hierarchical level for easier searching. The Yellow Pages, for instance, are full of (alphabetized) Categories, which are then alphabetized within each group. However, as an organizational technique, grouping is just as valid whether the group is ordered or not.

Location is the fifth 'hat rack', and this is the one that is really only the tip of the iceberg. Ordering and Grouping are both 'pure' arrangements, in that the data is compared only against other points in the set. Location is a specific form of what I would call 'mapping'. Mapping is different from the other two in that it cross references data from within the set, to an external dataset. Geographic location is one such possibility, but non-physical sets offer even more potential, particularly in an information-rich culture such as ours. Information states, methods of retrieval, and authorized user groups are just a few such sets. Computer network diagrams are often a hybrid of physical and informational mapping.

So my data closet has only 3, rather than 5, hat racks--though each has a multitude of different pegs. In principle, each can be combined and stacked with each other in limitless multiples, although for human parsing there is a practical limit. Simple tools, but--perhaps much like deciding which hat to wear--figuring out when to use which one is not always an easy task! Information Architects take this challenge as a full-time occupation.



Categories