Despite the complexity of the material Kalbach deals with throughout Chapter Two, he closes it by reiterating some concrete, practical advice. Three points emerge as being of immediate importance:

• Labelling is absolutely critical
• Organization must be sensible 
• Appearance plays an important role

Consistency in navigation is … important in  supporting the information seeking process. But too much consistency in navigation may be a bad thing. The notion of transitional volatility shows that variations to the navigation are actually important in creating a sense of movement through a web site. These changes, however, should be subtle and predictable.

Kalbach devotes a good part of Chapter Two to discussing the emotional aspects of information retrieval and web navigation. He then discusses a schema outlined by Carol Kuhlthau of Rutgers describing the typical arc of emotional responses in web searches.

The details of the schema are to my mind uninteresting, simply because they seem to assume a user performing a known-item search (interestingly, the concrete example Kalbach cites a couple of pages later purporting to apply Kuhlthau’s methods deviates from his earlier summary in significant ways). The overall message, however, is vital, and one that I think most web users will find familiar – a general emotional trajectory of optimism as a search begins; a sense of disappointment or frustration as the result set shows you that the task is not as simple as you thought it was; followed by growing confidence and optimism as you winnow these results down. The ideal, it seems, would be to minimize the sense of unhappiness felt in stage two – either by pre-winnowing the result set or, if this is impossible/undesirable, by some other means. It could presumably be possible to indicate that the breadth of material returned helps to solve the ‘not knowing what you need to know’ problem identified by Donna Maurer …

This article has got to be one of the most important and profound pieces on information design yet written. Unfortunately, it’s largely theoretical, meaning it’s not immediately obvious how to apply it to navigation solutions; furthermore, its chief point is that navigation is not strictly separable from content, meaning that there probably aren’t broad theoretical solutions for navigation anyway. But the issues it raises are complex, and need serious consideration.

Much of the article is concerned solely with attacking what was at the time of the article’s publication – 1997 – the dominant metaphor for web organization, physical navigation. Once the preliminary falderol is dispensed with, however, you get to the meat of the article – which is to say, the claim that the entire way we think about web navigation is wrong. In the conventional model, users employ navigation as a map to get to a particular screen displaying the information they’re looking for. According to Dillon and Vaughn, however, the processes of navigation and information can’t be disassociated in this way.

In Dillon and Vaughn’s account – backed up by some serious empirical evidence – information acquisition demands not just viewing of information, but that this information be assimilated into one’s own ‘knowledge structures’ and related to other information. In other words, content matters. Domain novices and domain experts, for instance, apparently use different means to orient themselves within documents. While physical organization (layout, site structure) may prove extremely helpful to novice users, expert users have complex rationales for orientation based on document content. Because they are familiar with the genre of particular writings, with the inference structure typical of these, and the verbal markers most often used for both of these, expert users employ comprehension of generic structure to find their way through documents.

It seems impossible to put the matter more succinctly than Dillon and Vaughn do themselves:

Comprehension is not something ‘other than’ navigation, some form of task that is independent of the process of moving through the information space. Rather it is an intrinsic component of information use … The purpose of moving through the information space is frequently the same purpose as the journey, to reach an end point of comprehension – and in this case the journey is the destination.

As noted above, the notion of information shape proposed as a means of describing this cognitive journey is a complex one ‘invok[ing] ideas of discourse communities, traditional forms and emerging digital document types’; there’s no magic bullet for this one.

The upside, however, is that this does open up the possibility that users might be guided through a site – that a site might in other words be ‘persuasive’ – in a natural and useful fashion arising out of the characteristics of digital genre; that is to say, navigation ’solutions’ will fall naturally out of content.

In a 2000 paper in First Monday, Choo, Detlor, and Turnbull attempt to outline a schematic classification of user behaviour based on information needs.

Informationally, this schema is based on four identified kinds of information acquisition, and six user behaviours that support these modes.

Choo et al. divide modes of information acquisition into,

1. Undirected viewing (more or less self-explanatory: this is ’surfing’)
2. Conditioned viewing (whereby ‘the individual directs viewing to information about selected topics or to certain types of information … The individual has isolated a number of areas of potential concern from undirected viewing, and is now sensitized to assess the significance of developments in those areas’)
3. Informal search (whereby ‘the individual actively looks for information to deepen the knowledge and understanding of a specific issue. It is informal in that it involves a relatively limited and unstructured effort’)
4. Formal search (whereby ‘the individual makes a deliberate or planned effort to obtain specific information or types of information about a particular issue. Search is formal because it is structured according to some pre-established procedure or methodology’ and the information sought is precise and intended to serve some particular, definable purpose, possibly in relation to policy)

Here’s another Flex annoyance – though I can see that from some users’ perspectives, this might be more of a feature than a bug.

The problem arose for me (as with most of my problems with Flex) in trying to apply a basically CSS/XHTML style workflow to a Flex application. I’m guessing most people who have done much fiddling about with CSS/XHTML design will have gotten used to creating graphics and texture files that are, say, maybe twice as large as they will generally need to be in order to accommodate different browsers and display resolutions: you can then size whatever component (probably a tab or a button) the graphic is for by percentage, confident that the background graphics file will be large enough for whatever platform the site visitor happens to be using. Other reasons for making an image larger than the area to be displayed might also include creating multiple states in a single file and then displaying them using positioning to save on bandwidth.

Unfortunately for this workflow, Flex has a peculiar behaviour when it comes to sizing and skins. Normally, it is possible to create a skin file larger than the component it skins – but only if the sizing is done in pixels (or rather, given that we’re talking about Flex, if the sizing is done without units). If the sizing is done using percentages, the component displays with the dimensions of the skin file instead.

I suppose a die-hard Flexhead would reply that one of the advantages of Flex and Flash is that you don’t have to worry as much about creating fluid layouts using these tools: you can set everything to be pixel-perfect, and the swf compiler does the magic for you. In addition, there’s no reason to be compiling excessively large graphics resources into the finished product. As with everything Flex, though, the (much appreciated!) close similarity to other web-technology workflows can lead to the assumption that they’re identical …

Donna Maurer’s short article ‘Four Modes of Information Seeking and How To Design For Them’  is a nice, short, practical summary of how users search for information on the web – although unfortunately, the two ‘modes’ she identifies as most under-catered for and most problematic are also those she has the fewest solutions for.

In brief, Maurer’s taxonomy of modes runs

1. Known item searching, whereby users ‘know what they want, know what words to use to describe it [and] may have a fairly good understanding of where to start’. Such searching is best served by:
   • Search boxes
   • A-Z indices
   • Quick links to frequently used items
   • Browsing by header
2. Exploratory searching, whereby site visitors ‘have some idea of what they need to know … however, they may or may not know how to articulate it and, if they can, may not yet know the right words to use. They will usually recognize when they have found the right answer, but may not know whether they have found enough information. In this mode, the information need will almost certainly change as they discover information and learn, and the gap between their current knowledge and their target knowledge narrows’. Effective searching in this mode favours,
   • Browsing, broadly defined
   • ‘Related information’ links
   • Search functions (although this is problematic, as part of the user’s difficulty might be lack of familiarity with keywords)
3. ‘Don’t know what you need to know’. In this mode, users aren’t aware of what they need to know; or may have an inaccurate idea of what they need to know; or may have no specific goal in mind in visiting a site. Maurer cites the following situations as archetypal examples of DNWYNTK problems:
   • Complex domains (legal, financial, or, one might add, study skills)
   • Areas in which the user is to be persuaded of the desirability (whether commercial, intellectual, ‘best practice’, etc.) of some course of action which may be counter-intuitive or involve extra effort
   • Unknown domains
   • Surfing in order to keep up to date, rather than to answer any particular question
     Maurer doesn’t offer many solutions here, but recommends
   • pages with straightforward answers to foreseeable questions (a bit like a glorified FAQ, one suspects, in most cases)
   • easy linking to find more detailed information
4. Refinding: this is an intriguing category, as I suspect almost anyone who spends much time on the web will feel s/he wastes a great deal of time merely hunting out previously found truffles. Despite the number of comments on the article to the contrary, furthermore, this is clearly not the same as known-item searching, as the examples of solutions given by Maurer indicate:
   • wishlists (a la Amazon)
   • Save For Later (a la eMusic)
   • favourites
   • del.icio.us

Looking over Toms’ article – admittedly written in 2001, when things looked a lot different – I’m underwhelmed. Toms nicely demonstrates that people parse documents in terms both of form and content, and that form is processed more quickly than is content. Given the speed with which people parse webpages, this is important. But Toms’ idea of what constitutes ‘genre’ seems curiously old-fashioned: she seems to be thinking predominantly in terms of layout and typography. This utterly fails to take into account,

1. the interactive nature of the web. Could it perhaps be argued that ‘genre’ on the web consists primarily of what a visitor can do on a site (in a button-pushing, strictly interactive sense), rather than what the site looks like? When I think of digital genres, I think not so much of “news sites” vs. “entertainment sites”, say, but in terms of wikis vs. e-commerce vs. blogs. The question of layout and typography might be more important in terms of alerting visitors to what is expected of them interactively rather than denoting what kind of information is found where on a page. 
2. the temporal nature of media generally. Toms appears to be thinking of the business letter of one or two pages as the archetypal document. But most documents are longer (or take more memory, to put it another way), and so are ‘navigated’ temporally as well as spatially: one doesn’t read a book or make a purchase on Amazon all on one page, and tools have to exist to guide and orient the user of either through the experience.

The notions of ‘information scent’ and ‘information shape’ are in Kalbach’s view both crucial to effective web navigation.

The concept of ’scent’ is (as the term is presumably meant to convey)  difficult to define concretely. Kalbach says it ‘refers to how well links and navigation match a visitor’s information need and how well they predict the content on the destination page’. He also, however, quotes Spool, Perfetti, and Brittan to the effect that scent is a quality compounded of many ingredients that adds up to a ’sense of confidence in navigating’.

Of particular interest in creating this sense of confidence is the powerful role of keywords and labels – users rely a great deal on the rapid scanning of a page for relevant labels in navigating, and this is furthermore an effect that increases as users become more experienced in using the web. Finding labels that are both accurate and map onto user expectation is thus crucial to creating effective web navigation.

Related to the concept of ’scent’ and confidence is the notion of ’satisficing’ – the observable behaviour whereby visitors will tend to choose the first option they encounter that seems to satisfy their needs, rather than evaluating an entire page for the optimal solution.

Complementing the notion of ’scent’ is the idea of ‘information shape’ – which is to say, information genre and the set of formal characteristics that define this. Because genres are such precisely in relation to other genres, overall rules are impossible to formulate: satisfying the conditions of information shape is really a question of working with established conventions. That being said, useful work has been done in the area of identifying evolving digital genres – with Kalbach citing the work of Elaine Toms and that of Andrew Dillon and Misha Vaughn in particular.

In a fascinating 2003 paper, David R. Danielson of Stanford University discusses web navigation in terms of ‘transitional volatility’ – that is to say, the extent to which pages change in both navigation display and content within a site.

In Danielson’s conceptual schema, visitors navigation and orientation is the result of three factors:

1. Habituation to a particular navigational and layout schema through exposure to a particular ‘patch’ of a website (this seems to correspond well to the ‘hub and spoke’ navigation pattern identified by Andrew Cockburn and Bruce McKenzie).
2. Prediction of the structure and nature of content based on this habituation when clicking on a link
3. Reorientation based on the content actually rendered

In Designing Web Navigation, Kalbach identifies six navigation ‘gotchas’ that decrease the credibility of a site:

1. Broad, vague categories, the meaning of which isn’t apparent to the user
2. Cryptic abbrevations
3. Poor organization of options within menus
4. Unexpected navigation behaviour (for instance, top-level links leading to non-HTML content)
5. Proof-reading errors