Archive for September, 2007

One thing that has crossed and stuck on my mind lately, is how information is presented on topic maps based web portals.

Topic Maps based portals typically, or should I say obviously, use a subject centric approach for presenting information: every page presents information about a subject (represented by a topic in the topic map).

And most topic maps based portals lay out the information about those subject in much the same way:

• Page name = topic name.
• Page body = occurrences; e.g. description or article text.
• Right column = associations.

(In addition to this, the left column will most likely consist the standard global navigation menu).

Let’s take a look at a few examples:

Screen shot of government.no.

Screen shot of bergen.kommune.no.

Screen shot of www.forskning.no.

Now, what’s interesting is that one of the most interesting parts of the topic map (as far as I consider), the part that enables associative navigation and makes a difference compared to other technologies; the associations, are placed in the right column — a common spot for banner ads and less-important/interesting information that users have learned to overlook.

This makes me wonder: has anyone ever done user testing on topic maps based portals that might reveal how many users use the associative feature of topic maps - excluding the use of menus and the like?

How many users figure that the right column holds interesting information about related subjects and actually click these links?

I doubt topic maps based portals are wide spread enough for users to know that the associations are of interest — and in any way, which user knows - or cares - what goes on behind the scenes? And does it even matter, as long as the site is usable (in general)?

I am pretty sure it is important (as always) to use effective visual cues in order to direct the user’s attention towards the associations. Some of the examples above do better at presenting the associations visually — forskning.no almost appears to be hiding them, while bergen.kommune.no’s approach doesn’t seem too bad — for user’s to spot them while scanning the page for useful information. Still, it makes me wonder …

Say, where did they put that “Home” (”Hjem”) button (the image without alt text)?

(www.sandnes-kulturhus.no)

In all fairness, it’s a public site by a cultural institution — which definitively has a limited web budget.
Still, it makes me wonder why they decided to put it at the bottom of the main menu.

Lars Marius Garshol has written an interesting piece about Search in Topic Maps Portals. I also want to point out a comment by Svein Ølnes, that refers to a very important issue:

I think the reason for Topic Maps based portals generally having better search is obvious: the search leans on an underlying semantic structure.
[...]
The problem with some Topic Maps based search facilities is the tempting “let’s show them all we know”. The key to good searching is to restrict the features to the most useful and not try to show all the information in the Topic Map.

Without doubt, this is a major problem at some Topic Maps based portals. It might be tempting to think that users care about being served all the nicely structured information simply because you can, but too much information may be confusing to users whose prerequisites are not the same as the Topic Maps experts’.

After all, when you search Google - something most user’s are familiar with - the result is fairly easy to interpret - the most relevant result is presented at the top of the list. And that’s it.

If only JavaScript developers would stop using alert() for debugging - at least while in production. Searching for flight tickets on www.berg-hansen.no gave me the following:

The page at http://www.solfaktor.com/ says:
TXL not found.

Now, there’s something every user needs to know.

See also: Web 2.0 Usability.