This week, I read chapters five and six of Information Architecture for the World Wide Web. This book is so well-structured and "common sense," I find myself nodding and agreeing with practically all of the content.
Chapter five covers labeling systems. The focus is on determining an entire system, rather than producing individual labels. Types of systems include navigation labels, indexing/meta tags, links, titles/heading, and icons. (Obviously,) labels should follow the organization scheme. To pick labels, you can look at old version of the site, competitors' sites, controlled vocabularies, the content (such as document titles), or poll users themselves. Once you've devised a system, you should list all your labels together and examine them as a complete system in one place. I thought this was a particularly good idea. The point is to have extremely high consistency across the site. Also, your labels should be as clear and unambiguous to your users as possible.
One pondering I had is whether ease of labeling a category is related at all to its prototype/diametric nature. I mean, if a category is well-defined based on a limited number of attributes, then it seems labels could be defined from these attributes. Whereas, if the perception of the category is more of a gestalt thing, then labels may be harder. I imagined an e-commerce site with a label: "Things we sell." (Not very suave, but it works for this example. Maybe "Products" would be more mainstream.) Now this would be all physical items that the company sells. Pretty well-defined. But what about warranties or on-site maintenance? These aren't physical, but likely very closely related to individual products. A prototype-ish example would be trying to label archived threads of discussion board messages. In this case, most of the messages are probably close to a main topic, but there are very likely outliers and other subjects mentioned. Would this be harder to label? (It's also hard to think about these things -- single labels -- without the context of the other alternate categories. Hence, the push for labeling systems I guess.)
Another thought I had is: how necessary is a label to understanding a category? This is basically the old question of whether you can think about something without being able to name it. I imagine it's easier with a label. Yet would it ever be possible to demonstrate site organization without the use of labels?
Chapter six gets into searching systems. It seems that a large percentage of people search as their primary (sometimes only) mode of navigation. But a site search engine should not serve as a bandage for poor site structure. There are different types of searching, such as known item (one correct answer), existence (does this item exist?), exploratory (tell me more/what's out there), and comprehensive (give me everything on this topic). Depending on your users, you need to customize the interface, whether keeping it simple or including more options. What options are available need to be clear. Also, choices need to be made on how to display results or how much to allow users to customize this (indeed, do your users care enough to bother customizing?). It is also important to consider what should be indexed. Generally, destinations pages should be indexed, while navigation pages should not. (Except, I think, when searching can be used to find the appropriate nagivational category; how would it be best to combine these two without swamping the index?) Multiple subject, content, or audience specific indices can be a nice feature. (Indeed, limiting a search to a certain branch of a site is a great way to combine browsing and searching.) One thing mentioned in this chapter was granting the user the ability to both see how relevance is determine and even to be able to change this. I have never encountered that before. A good idea!
I also finished The Fine Line this week. This book was not all that great. The basic conclusion, which was clear at the end, was that humans seem to be of two minds: the rigid mind, in which categories have very clear boundaries and ambiguity is discouraged; and the fuzzy mind, in which categories are very blurry and fade into each other. It seems that the rigid mind fits in with the logical (what I often call diametric) view of categories, while the fuzzy mind describes graded and prototype categories. The author also proposes that the fuzzy mind corresponds with the nature of the world and primitive or early experiences, while rigid mind is more of a product of logical human thought and abstraction. I notice that this correlates with my thought last week that abstract categories may be more clear and easily defined. Throughout this book, the author draws connections between a great number of human activities and whether these activities are dictated by the rigid mind or the fuzzy mind.
To Week 5 →
LIS: Week 4 -- Readings and Comments
|Last Edited: 26 Sep 2001|
©2001 by Z. Tomaszewski.