Effective View Navigation
If I were to edit a historical collection of essays on information architecture, it would have to include George Furnas' Effective View Navigation (PDF). Published in 1997 and borrowing from earlier work that harks back to '95, it is not to be read for new methods of navigation. It does however provide serious analysis of a fundamental information architecture challenge, 'despite the vastness of an information structure, the views must be small, moving around must not take too many steps and the route to any target be must be discoverable.' His focus is on what to do in that small view (i.e. a page, or a viewport of a page) including ideas that form the basis of information scent.
For serious information architects, EVN is a must read. At eight ACM-style pages it's not long but it is dense, requiring me to double-back often. His initial discussion of Efficient View Traversibility sets aside the user experience perspective in favor of graph theory and so can be a dense read to the practitioner, but it is necessary to his later arguments which are worth understanding.
Here come the spoilers: after an interesting logical analysis of the problem he concedes that one navigation scheme - what we might call global navigation today - would be hard-pressed to provide access to a large information structure. Along the way he predicts the rise of a combination of global and local navigation.
His ideas of efficient view traversibility provide (in my interpretation) a good mathematical explanation of creating short paths to target information and a helpful last resort after other methods fail, 'always remember the strategy of putting a traversable infrastructure on an otherwise unruly information structure!' For example, as a last resort, stuff the information into a balanced tree. If we combine this with a user-centered design method, we might say, 'If research of the users does not reveal a clear path to the information, and the users understand the domain enough to understand the meaning of a set of hierarchical categories, then stuff the information into a balanced tree.' Other conclusions of his, combined with a design process, can yield a more systematic design method than what we have now (which is why I'm trudging through this stuff).
His use of navigation requirements nicely frames his discussion, for example, 'Every node must have good residue at every other node.' We rarely have such strict requirements in our designs, but the practice of using navigation requirements - falling somewhere between scenarios and screen designs - helps increase the likelihood of creating successful navigation.
His setting aside of user experience considerations at times seems unworkable; the goal of efficiency can be at odds with what may be cognitively effective with users. But the thoughtful designer can factor that into the use of his ideas.