Humans remember things in a fuzzy way. The connections we make between disparate objects happen because of what they share in common – in time, in space, in smell and sound. Burning leaves? Grandparents’ house. Trampoline? 8th grade pool party. Florida? Hanging chads.

I’ve read about techniques of running through your rolodex, or address book, occasionally and without reason, just to see the names and visualize each person. This has been shown to be very effective in helping to remember people’s names since we are more likely to remember what we have seen recently – and forget what we have not.

Reinforcement plays a large role in memory and recall.

If websites begin to implement a tag cloud epoch, they can begin to “forget” the stale tags in their system. Sites can begin to have their most visible tags dissolve in an organic, human way. As people do not continue to tag a certain thing a certain way, this thing should fade slowly from view. It should still be findable (and re-findable) through search and browse, but the tags describing it should count less and less when considering what hot lists to put the item on.

This arbitrary date of oldness, this epoch, should be customizable, of course. But it should be available and it should be prominent. Allow a user to define how far back the tags should be counted. Allow a user to define how old is too old and how recent is recent enough.

Folksonomy will be around for a while, I suspect. It has proven itself useful in many ways – probably some yet to be seen. Some of yesterday’s simple tags will seem quaint tomorrow. Give the user the power to decide whether quaint is signal or quaint is noise. Go implement the “Since” date filter on all your tag clouds today.

Tag clouds are young. The term folksonomy itself was coined just over a year ago by Thomas Vander Wal on the IA Institute’s Members’ Discussion List in a response to Gene Smith and has yet to be a word even a medium-sized minority has heard, much less come to understand.

However, here on Cloudalicio.us, we can already see that many items being tagged since the beginning of del.icio.us have stabilized. New tags being added to these URLs are no longer visibly affecting the overall weights as each tag’s marginal weight has decreased since the first tag was assigned.

We have a better view of what everyone “has thought” of an item by looking at it’s full, cumulative tag cloud, but we miss what everyone “is thinking” of an item when we allow ourselves to look back into time and see possibly “stale” tags from long ago intermingled with more contemporary, more recent tags.

This is not to say that tags from long ago are not valid or no longer important. They are. It’s just that if we’re going to be looking at tag clouds as tea leaves into what a population thinks about something, we need to recognize that unless we silence the voices from long ago, they’ll continue to speak to us and influence our counts. Opinion polls for presidential elections are considered old almost before they get to press. Yesterday’s numbers don’t mean anything. Having snapshots in time is what allows trending to be done.

Trending in tags, here on Cloudalicio.us and everywhere else, has been cumulative only. We do not yet have the tools to slice all this new data and make it more meaningful than to look at it in toto, in summary form.

And so, we need to define a tag cloud epoch.

Tag clouds should be viewed within a framework of five points in time.

1. system epoch
2. tag cloud epoch
3. zoom window start
4. zoom window end
5. now

As it currently stands, most systems that have tagging enabled count all the tags between the two outermost points in time (#1 and #5). Cloudalicio.us currently allows four of these to be defined (#1, #3, #4, #5) when used with Date Zooming. Unfortunately, there is no current way to define the point in time from which tags should start to be counted. The addition of point #2 is what will let us see any arbitrary slice in time. This will allow us to see what a population was thinking during a specific week or month about a specific item.

Hopefully, it will appear here soon…