AskTog: Interaction Design Solutions for the Real World
Interaction Design Section   Living Section   About Bruce Tognazzini
NN/g Home > AskTog > Columns > November, 1998, Reader Mail Ask Tog, November, 1998

Reader Mail

This month's mail:

An Alternative to Website Navigation Bars

I found Website Navigation Bars very interesting, but my enthusiasm for the multilevel solution (Stu's no. 3) is dampened somewhat by most users' general lack of appreciation for hierarchical menus. We need to flatten the menu -- both in terms of space used, but also to lighten the users' burden in building a mental model.

One technique that I have played around with in an attempt to reduce complexity is to use a drop-down menu list that has an indented outline. For example:


Picking Item 1 will bring the user to a summary page which would contain links to SubItem 1 etc. If you want to see an attempt to make this works in practice (and maybe find an interesting food recipe!), point your (DHTML enabled) browser to My implementation fails to address the mapping problem—it wasn't intended to. Still, it might trigger some ideas.

John Lloyd-Jones

Stu's solution is more complex, but that is because it is defining a more complex space. I agree that Job One is always to create as simple a space as possible. Creating a simple hierarchy, however, is not always achievable, particularly on large, complex sites. When such complexity does arise, it is easy to decide the best way to handle it is to hide it. Not true. It is far better to reveal it through a map, no matter how complex the map must be. At the very least, it will bring the problem into the open, and those responsible for the complexity may very well decide they really can live with greater simplicity, even at the expense of a few features.

When you can keep it simple, do. John's solution is elegant.

Representing link destinations

In last month's Sorry State of the Web, you failed to fill us in on an important point. Do you have recommendations for distinguishing between intra-page, inter-page, and other-site links?

At one point I was using italics for other-site (external) links. Zeldman opens new windows for external links; I don't like that because naive users who have their browser window maximized all of a sudden find the "back" button grayed out.

CNN always puts a disclaimer on links that go to external sites, saying that CNN isn't responsible for the content, although I think that's more CYA (Cover Your Assets) than Human Interface.

J. David Eisenberg

Isn't it ironic that the lawyers are doing a better design job than we are? The lack of a standard appearance for the three levels of destination is a major flaw in the HTML spec. The eventual cure must come from those with control over that spec.

In the meantime, we are left with a vacuum. No individual web site can fill it. There doesn't seem to be any natural text style that will intrinsically communicate distance. Instead it will have to be learned, and learning of such things only really works if a large percentage of sites are doing it. Individual websites will, until this is straightened out, need to depend on explicit references, using lots of words or really great icons.

So how is the change ultimately going to occur? It used to be that Netscape, Microsoft, and the standards committees held all power. However, we do have a new player--the makers of site editors. Were GoLive, Macromedia, Adobe, or some other premiere site-editor manufacturer to adopt a standard that would automatically style links depending on their destination, others would, with luck, soon follow suit and the standard would be established. Guys, are you listening?


(Click for continuing discussion.)

Protecting your user's information

Just read Silo Design for Web Transactions. Some good points. Another one to add to the list is "never, ever, lose a person's info."

I've been advising my brother, who is putting together a small site (Why? To sell hats, of course.) and one trick he's relying on for the commerce part of things is that a person viewing the site can pop out of their current location in the site and go somewhere else without losing any of the information they've entered so far. If they forget that they need a hat for Aunt Sadie in the middle of entering the gift card text for Cousin Jim, they can abandon the card, browse some more to find Aunt Sadie the perfect toque, all the while not losing the shipping info for Cousin Jim.

It requires a modicum of database smarts and a solid handle on how cookies/session IDs work, but it's an amazingly freeing design concept for a transaction/commerce site. (Making sure your site hierarchy is at most two layers deep helps, too!)

Rick Levine
Sun Microsystems

Rick's bro is doing it right. Users should never lose their work, no matter how much work guaranteeing that causes us.

Web pages are stateless. When many folks read that back at the beginning of this revolution, they took it to mean that their sites could be stateless, no matter what kind of mischief it caused. What it really means is that we take on the obligation to construct by hand what those www guys left out. We have not been freed from an obligation to protect our users by maintaining state on our own.

Simplicity is Disappearing

I was working with a programmer recently to create an application that Emergency Department nurses would use to communicate the need for a bed to the custodians of a distant patient care unit. We drew out a workflow diagram and began to think about a fundamental database design to support the facility.

As we talked I realized the programmer was relating everything I said to the most esoteric facilities in VB6 as a solution to first one programming challenge and then another. This sounded familiar. I'm not a programmer but I do know a few things about software design.

As my programmer pondered how to structure a complex SQL query to provide a simple on screen report I realized that he was overlooking the obvious. In our case a simple filtered sort would fill the bill. The programmer was astonished. How could he have overlooked this simple approach to what had appeared to be so complex a problem?

More often than not, I find that the complexity of software is directly proportional to the complexity of the tools available to program it. Perhaps a return to a more simplistic approach will result in tighter and more robust applications that actually perform as intended. Maybe next year.

Charles Fox

I think you are being too hard on the programmer. The genius of simplicity tends to arise from intuitive leap. Once one of these occur, all previous solutions tend to look strained and baroque.

My experience has been that the simplest of interfaces are only possible within complex environments. It is easily an order of magnitude more difficult to create a simple pure HTML interface than a simple visual BASIC interface, just because HTML as a tool is so pitifully weak.

Yet, I agree with you completely about the complexity of the tools. I refer, however, to the interface between programmer and tool, rather than the capabilities of tool itself.

When the interface to a tool is weird and ugly, two things occur. First, and perhaps most importantly, that tool selects for people who can learn to use it. Someone who can use C++ obviously has superior abilities to the average person on the street. Unless programmers are self-aware—and most are not—they may not recognize those superior abilities, and their solutions will tend to assume everyone has similar abilities to their own.

Second, weird and ugly tools teach bad habits, likeWritingThingsLikeThis, instead of like this. When I worked on development tools at Apple, I put high priority on reflecting in the development tools the same interface I wanted Macintosh users to experience. Visual BASIC has actually taken a similar path, offering layout tools that are very Windows-like, reserving separate, special places for things taht exist only within the programmer's world, such as code segments. The lack of a true object model, however, requires programmers to memorize complex models to be able to find their way around their own programs, a particular branch of weird and ugly that far too often will passed down to their users.

Click for a reader response

Macintosh survival guide


I enjoyed your presentation(s) at the UIE conference this month in Cambridge.

My question is about Apple, not about design, so don't feel obligated to respond. My hope is that your experience might answer a business question and help me predict some things, though.

In addition to web site managment, I support 25 Mac users. All around me in this building are PC MIS departments who are very near to winning the hearts, minds and checkbooks of our senior managers and forcing everyone to use a PC. Well, nearly so, anyway. Frankly, they've got an easier selling job than they used to have.

A year ago I was sadly preparing to move myself and my users there, too. I can read the signs as well as anyone else, even though I always have had a reverence for the underdog. However, the past year's resurgence of Apple into the consumer market has me hopeful again. I don't believe they've done a good job of modernizing the OS (until Rhapsody/OSX, that is), but the job Jobs seems to be engaged in is keeping Apple alive long enough to migrate to a modern OS.

My question: What's your educated guess about whether they'll prosper enough to take the pressure off people like me? Am I engaging in wishful thinking to assume that we'll still have Apple to kick around a year from now? Two years?

Again, thanks for the presentation and advice.

Doug Stanfield

A) Apple will prosper. B) It won't take the pressure off you. MIS departments are conservative to the extreme. They used to shy away from anything that didn't say IBM. Since IBM handed their leadership position to Microsoft on a silver platter, these same MIS people now shy away from anything that doesn't say Microsoft. That will not change in the foreseeable future.

Fortunately, in spite of Apple's best efforts, you can today easily integrate your Apple machines thoroughly into the Windows world using two inexpensive tools. The first tool is Dave. This lovely bit of software almost effortlessly turns your Mac into a first-class Windows client. People can see you on the Windows network, and, through the Macintosh Chooser in the usual way, you can see them.

Second, Quickeys for the Mac will help keep your sanity if you are one of those, like me, who must dance back and forth between Windows and Macintosh machines. Just redefine your favorite Windows short-cut keys—Control-Z, -X, -C, -V, -A, etc.—to map to the same keys on the Mac. That way, you don't hve to switch mental gears every time you move from machine to machine.

Help Wanted in Merry Olde England

Why aren't there more HCI jobs out there? I graduated from an HCI course 3 years ago. I have yet to practice my skills full-time. The best I can do is create a niche for myself in the area whilst doing other work...doing usability as a side skill with my main duties being producing multimedia or educational software, programming, or some such.

I am gradually seeing products becoming more usable (slightly) and the public seem to be demanding this more, but no company seems to want HCI graduates. Do they know we exist? Is there a market for us?

In the US the demand seems to be greater, but here in England if you do a search for HCI jobs you get zip, zilch, nothing!

What's going on?


Perhaps some of my readers across the pond can give us some insights. HCI professionals are hot here in the Valley. I, frankly, wasn't aware that England was lagging so far behind. What are things like in other parts of Europe and the rest of the world? What can be done to help out? And if you have a job for Stu, let me know. I'll pass it along to him.

First Response to Help Wanted

The UK is a usability desert ! Which is a shame because we have had a lot of really good basic research going on for years.

I suspect the lack of application is to do with the "immaturity" of the software producers here. My impression is that the Scandinavian countries are the only holders of the HCI torch in Europe for the time being.

Nathan Langley

How about adding discussions to AskTog?

Hi tog,

Your pages are looking better and better. It is now very easy to distinguish between the current articles and the list of previous articles. The articles themselves are thought provoking and well focused and give actual design recommendations, which is great. In fact, you seem to be following the guidelines in the 'how to publish a great user manual' article.

One area I especially like is the 'ask tog' section. It starts to make your pages feel alive, with some two-way interaction going on. Might I suggest this is expanded with the addition of some kind of discussion forum or chat area. This could be used for people to have their say on the burning issues of the month, or questions that you pose. You could feature the most interesting responses (is this beginning to sound like too much work???).

Anyway, keep up the good work!
Stu - the perpetually frustrated HCI guy in england

I spent a lot of time exploring adding discussions. In the end, it was going to cost way too much money for the web server space for a zero-profit zine like AskTog.


Don't miss the next action-packed column!
Receive a brief notice when new columns are posted by sending a blank email to

return to top

Contact Us:  Bruce Tognazzini
Copyright Bruce Tognazzini.  All Rights Reserved