Keeping the previous structure intact is an vital commitment for information architects. Moving forward is required for success, but maintaining continuity with the past is essential to maintaining balance. This concept is rather broad, but can be illustrated in a few examples.

Imagine, for example, you have a website which sells hundreds of products that are poorly organized. Let us say you sell books. You have a category for “Romance,” another for “Renaissance Literature,” another for “Robert Louis Stevenson,” and yet another for “Hardbacks.” What is wrong with this picture?

Bascially, the main issue here is that a static category structure (if one must be used at all) should remain as simple as possible. Specifically, any category structure must be centered about one and only one attribute. Described above are category names which reference four different types of attributes: Genre, Time, Author, Hardback/Paperback.

The short solution is to reduce the category scheme to one attribute, and then delegate the remaining category attributes to the product level. In other words, if Genre was decided upon as the key attribute, and it was decided that an edition of Remembrance of Things Past should be placed into it, certain field values would be associated with this product to compensate for the lost categories. Time=19th century, Author= Marcel Proust, Cover=Hardback, etc.

From this point, with some basic coding, the user should be able to “drill down” to the exact, desired product will little effort, only by knowing beforehand the genre he/she desired.

How then, do we implement a new, more intelligent structure, while keeping the previous structure intact? There are a number of techniques involved here. But you can first be sure that any old urls used in your site (for example, if your site has high ranking pages indexed in the search engines, or links from other high traffic web sites) will remain active following the transition. The goal is not to kill the old urls immediately, but to phase them out gradually, at a natural pace, as the new structure takes effect.

In order to learn more about how Informationarchitech can clean up your website categories, while keeping the previous structure intact, click here to request a free consultation.

The current development of search technology is demonstrating a much greater awareness of intelligence and meaning in information than we have had before. As chaos theory has taught us, the multiplication of virtually (and perhaps literally) any set of data along enough iterations will render certain patterns visible, and there are even patterns to the patterns.

What the algorithm developers at Google are only beginning to understand is that certain sets of rules can effectively lead users through the pathways of these basically natural patterns, which are discovered best by observation, experience and long thought. In so doing, they have developed a search technology that is so effective that it manages to help billions of users sort through the single largest collection of data in human history without the use of category, hierarchy, geography, or chronology.

These are, by the way, the five techniques posed by Richard Wurman as the principle means by which humans order information (the other was alphabetization, which to me seems fundamentally the same as chronology… what I call ‘chronology’ was separated by Wurman into “time” and “alphabetization.”)

But of these techniques, where does Google fall? Clearly, none of these terms even come close to describing what Google does.

How might one describe Google’s technique? Matching? A variable weighting system? A highly accelarated system of trial and error? It is difficult to classify for the same reason, I would venture to guess, that Wurman did not think to include it. The fact is, what Google is doing is a very new thing, because it is operating in an environment with which we still have little experience. Put simply, the dynamic rather than the static.

Of all of the techniques mentioned by Wurman all but “time” are basically static in human terms, and this only because time itself is always in motion. Categories have traditionally sliced through data, whereas Google literally “crawls” its way around it. The data is completely without a static state. It has no structure to speak of. No doubt the pages indexed by Google are entered into their own database in some kind of chronological order, but this is at most a key field, basically an index of the physical location of the data on the hardware. As far as any user is concerned, all of the pages are laid out at random, as all of the pages of a thousand encyclopedias placed upon a surface as large as the moon.

What is returned, after entering any term of their spontaneous choosing into a blank search box, is a work of magic—a rabbit pulled from a hat. It is based upon the idea of letting the information order itself. Information is, after all, a product of human intelligence. This means that the seed of intelligence is contained within all information. Intelligence is dynamic, and so should be the pathways used to find the information it generates.

While Google has delievered incredible “results” in a couple of senses, I believe the technique will reach its upper limit fairly soon. They can only keep modifying the search algorithm so long. As information continues to explode (many new sites created every day, very few deleted) so will Google’s infrastructure begin to experience stress. It will need to find another way to break the information down. But moving only forward! Attempts to slice the information up again into static divisions declared from above will fail, if they are ever made. They will either be too broad to have any meaning, or so narrow as to accelarate a “stupidification” process.

There is order to the information that is more than just pulling a new rabbit out of the same hat every time. But the nature of that order changes for every person on every day. There are categories, hierarchies, geographies and chronologies which make perfect sense for most all of the data indexed by Google. But they will never be the same for every person. I believe Google will (and if they do not, some one else will) find a clean method by which to combine categorization with dynamic search results. But the categories will be dynamic, not static. They will not simply change according to the needs of a given situation—the situation itself will create them.

Categories will appear in the the same way that pages currently appear—on the fly, as though pulled out of a hat. But unlike a page result—which is basically an endpoint, and either is or is not what you were looking for—the dynamic, or virtual categories will be forks in the road. They will appear depending upon the general direction you are headed, as well as what surrounds you at a given point in time. (always up to date) Most importantly, they will be circular rather hierarchical. They will be true pathways, making a full circle possible.

“The beginning of all understanding is classification.” -Hayden White

Categories are, first and foremost, a human invention. There are no categories in the natural world. Even the worlds we create artificially evade classification. The more complex any collection of information, the more difficult it becomes to invent a structure of categories which can describe the data without contradiction, ambiguity or misrepresentation.

Those who have little experience with the subject dismiss it immediately. After all, how hard can it be to divide information into logical categories? Especially when you already know everything there is to know about the products, services or pieces of information you are trying to organize! Nonetheless, anyone who has tried to sit down and order data in a logical manner has seen the problem all to clearly. For any given set of data, there are multiple ways in which it can be organized, each one as logical as the other.

The trouble comes when it is believed that one scheme must be placed over the other; when one attribute must be considered more important than any other. Informationarchitech refuses to adhere to strict dichtomies, in the belief that information is infinitely malleable, and like meaning itself, is forever in flux.

Static categories are out of date, a leave-over from the brick-and-mortar days in which we had to place particular items in separate aisles of even the most general of stores. Now we must begin to understand virtual categories, which are created or dropped in an instant’s notice, according to the needs of a given moment. Whether you have an existing category structure that needs significant pruning, or have just been bombarded with a set of data that you have no idea how to order, Informationarchitech is here to help.

Note: This is a very old post that appeared on this website a few years ago, several layouts and two CMS’s past! The examples in this article do not work, due to the huge amount of effort required in getting all the old (pre-Jquery) javascript to work again. I may update this article one day using modern technology/libraries.

It is amazing how quickly time (and technology) flies. It was not so long ago that web developers were blown away by CSS drop down menus that appeared in a flash whenever you hovered over a button in the header. Websites like A List Apart taught us how the hover property could be combined with list styling to evoke interesting effects. It wasn’t long before any website without drop down menus, or at least some hover-triggered effect, seemed boring and out-of-date.

We have learned a lot since then. In particular, the popularized “discovery” of AJAX showed us how the interaction of javascript, CSS and back-end scripting could forever change the way we look at a web page. Alongside this, the demand for even smoother effects is on the rise. Collapsable/expandable directory trees, smooth sliding tabs, and other features are becoming coveted by owners of website companies and individuals alike who want their website to be clearly on the “polished edge” of Internet technology.

Like so many “discoveries” in the world of technology these developments are not “new” so much as they have just recently registered on the public’s consciousness. The bottom line is people want them, and as developers, we need to know how to give them what they want.

One interesting effect that can be integrated into a navigational menu is the evolutionary descendent of the CSS drop down menu: the sliding navigational menu. The sliding navigational tab can take many forms, of course. But the challenge of creating “well behaved” as well as “complaint” smooth menu effects is a major issue for those who have just discovered AJAX or DHTML and want to integrate it into this design.

What are you talking about?

A picture is worth a thousand words, and a working example is worth even more. Hover over the tabs below to get an example of what is meant by “sliding navigational menus.”

In an attempt to help other developers who want to integrate smooth navigational features like sliding tabs into their website while remaining friendly to both user experience and search engines, I offer this two part series.

Part 1 explains the general considerations one should take into account when developing a sliding navigational scheme, and is intended to be somewhat understandable by those with little to no programming experience.

Part 2 is a step by step tutorial on the implementation of sliding tabs, as shown in the example above, and goes into some detail about the how to address the issues raised in Part 1 with particular coding techniques.

Part 1 - General considerations for sliding navigational menus. What a difference half a second makes!

Sliding tabs: what you want, and what you do not want.

What you do not want are confusing menus that open when you do not want them to, close when you want them to stay open, or once opened, stay open forever. You definitely do not want menus that get in your users way, or otherwise impede their search for your information, products or services. You do not want to bury important content in the menus that cannot be found by other means.

What you want are smooth, unobtrusive menus that function intuitively for the user, so that no explanation is necessary. You want them to be logically placed, and contain content that makes sense given the overall layout of your page. For users that do not have javascript enabled, you want to give your users an alternately and equally intuitive method for finding the information. You want all relevant links in the menus to be readable by the search engines, and all irrelevant links to be dynamically loaded upon the user’s request.1 (1-see Hide and Seek: the dilemma of dynamic content…”)

What you want are menus that look right, feel right, and like any navigational element, should help your user find your information more quickly than they could if it were not there.

Anatomy of a “slide”

Suppose, as in the example above, you have 3 different navigational links lined up next to one another.

There are essentially four events that can take place in your menu:

• User hovers over tab
• User moves away from tab
• User hovers over menu
• User moves away from menu

The user of course can move the mouse cursor very quickly across the screen, meaning you have to account for many different intersections of these events. What if, for example, the user hovers over all of the tabs very quickly? Will they all open and stay open? Will they all open and close one at a time? You must really think about what will make the most sense from the users perspective.

Here are two examples of the way a sliding menu might behave. One is “bad behavior,” and one is “good behavior.” Can you tell which is which?

Example 1.

Example 2.

It is impossible to cover all of the possible combinations of “events” and “states,” so let us take a few examples.

User hovers over the tab.

This seems straightforward at first glance. After all, if a user hovers over the tab, you just open it, right?

Not necessarily. First, how can you be sure the user really wants to open the tab? Perhaps the user was just moving the mouse to the top of the screen so they could, for example, view their shopping cart? If you immediately open a menu when the tab is hovered over, especially if the tab is at the top of the screen, you could annoy the user at best, and obstruct the user at worst.

A good solution to this problem is to insert a half-second delay before taking any action. It seems like very little, but it makes a huge difference. Delaying the action by half a second puts things in a more human time-frame, and “intentionality” is established before anything happens. If the user simply scans over the tabs quickly, none of them open.

Once the delay has transpired, you must determine if a menu is already open. If a different menu other then the one they have “intentionally” hovered over is open, the other must first be closed before this one can be opened. If it is the same menu, nothing need be done.

If no menu is open, of course, then the appropriate menu will be opened at that time.

User moves away from the tab.

The same delay should be applied when the user moves away from the tab as well, for at least two reasons. One, the user may accidentally move the cursor slightly away from the tab for a brief instant, but does not intend to close it. Two, the user may decide to move the cursor into the menu portion. If the menu closes the instant the user leaves the tab, they will not have the option to select anything on the menu.

In other words, you start a timer when the user leaves the tab. If no other events are triggered after this point, the tab simply closes. (Or if it is closed, nothing happens at all.) If, however, the user moves into the menu area, another function stops the timer that started when the user left the tab area.

User hovers over menu.

This event is simple. If the user hovers over the menu, it is clearly open and if you have done all of your programming correctly, you know no other menu is open either. All you must do in this case is stop the timer that started when the user left the tab. They have moved into the menu area, so clearly they want it open.

User hovers away from menu.

This is the same series of actions that take place when a user moves away from a tab. Did they really mean to leave the tab? A half-second delay should determine that. If the user moves back into the menu area (or the tab area) all timers are stopped and the tab stays open.

Timers are on our side.

Clearly, half second time delays can dramatically enhance the “usability” of smooth, dynamic navigational structures such as sliding menus. Half second delays both establish intentionality, and allow room for error. The motion of your elements becomes much more organic and sensitive to the user experience by allowing a half second “are you sure?” delay before changing the state of any element on the page. The user, of course, will not be conscious of this delay. If it were not there, however, something simply would not “feel” right, nor would the menus operate in a way that made sense to the user.