The problem

I’ve been considering how to quickly and efficiently display a long list of 400+ links for a site in progress. I’m using the word “link” here but really they are well-defined tags that people can easily classify their content by using just one of these terms.

Some ideas

While the links could be organised into subcategories and categories, the categories themselves would often be a hinderance as different people would classify the links in different ways or young people may not know or have thought about the category for the link ever (e.g. if the links were about books, the word non-fiction is not well understood). Having users make a series of decisions about the categories before seeing the links could be confusing and time-consuming with multiple clicks.

The links could be put into a massive tag cloud, i.e. ordered alphabetically and sized relative to something (e.g. number of items tagged). When doing a test of this with 400+ links it became pretty hard to quickly scan through and find the link you’re looking for.

The links could be put into “clustered” tag clouds as described by Hassan-Monteroa and Herrero-Solanaa in their upcoming paper. This puts similar tags together on a line as a cluster and puts similar clusters together vertically. However, for this list of links, the clusters would be quite big (long lines) and subjective (similar to the first problem of classifying the links).

I’ve then thought about a tried-and-true method: just listing the links alphabetically, all 400+ of them. Scanning down the list of links is much easier than in a tag cloud but of course this takes up a lot more room and it’s a little endless. I then chose to split the links up into more scannable groups (A-G, H-M and N-Z) and arrange these groups into four nicely spaced columns with group headings. Bringing back in the tag cloud concept, I’ve made the links different sizes, depending on the number of items tagged with that tag to potentially add extra help while scanning down the list of links.

A new (?) technique

To further (hopefully) aid scanning of these links, I am using what I believe to be a new technique I’ve had in mind for some time (but for a different application). While color or color shade has been used as another reinforcer of the popularity of the tags (e.g. popular links are large and black, the less popular the tag, the smaller and lighter grey they become), I had a different idea in mind.

Colors or shades could be used to display personal preferences, rather than community-aggregated preferences. In other words, tags I’m interested in or like or use or look at frequently myself (any of these could be implemented) would be a different color or shade to ones I’m not interested in or dislike or never look at. At a glance, I could see how my preferences/useage compares with the community’s as a whole.

I’ll call this a personal+community tag cloud where color is personal and size is community related.

So, to conclude, I’m going to be displaying my list of 400+ links in four columns, sized by community preferences and colored by personal preference. It’s not quite a personal+community tag cloud but is based on this.

When this (large) site goes live with this implementation (sometime before Christmas I expect) I will let you know!