The silver lining to the cloud of increasing search awfulness is that it’s forced me to think deeply about what search queries are. This has lead me to consider the idea of topical knowledge as an atomistic concept, an ever-shifting cloud of ideas attached to a central notion. The central notion can be as general as “history” or as specific as a particular sub-species of moss.
The cloud of ideas will vary depending on the context you apply to the notion. For example, you might want to know about a species of moss and its use in traditional knowledge. That’s one context. Or you might want to know about its relationship to its ecosystem. That’s another context. Each of these contexts will have a different set of ideas with only some overlap in this case.
Okay, so every topic has a variable set of concepts attached to it depending on the applied context. So what? First, of course, it means you can make great searches if you take your query terms out of the same concept sets. But it also means that you can use the data intrinsic to each concept to mold your search in interesting ways.
Consider another moss-related context: the use of a moss species in nature-based solutions (a method of response to climate change and other ecological issues). Nature-based solutions, as a term, was not used until the late 2000s. Therefore if you used this phrase in a Web search along with your original moss topic, your results would naturally limit themselves to the late 2000s forward. Nature-based solutions has a temporal relevance to the topic of moss, in this case because the term wasn’t invented until relatively recently — the concept had no temporal relevance to the topic of moss because the concept didn’t exist!
If you had a way to identify the temporal relevance of each concept around a central notion and map it to a timeline for that notion, you’d be able to develop searches that focused your results on a particular time-period without enacting a date-search method. Four-dimensional search queries!
I believe as web sites get more complicated and dynamic content continues to cause problems in traditional date search, it will behoove us to consider the idea of date-based browsing or searching in a more granular way.
I’m experimenting with Wikipedia to grab articles at specified intervals, develop concept sets from each iteration, and then do some filtering and analysis. The end result (I hope) will be a timeline with clusters of associated concepts which I can use to bend my search results to a specific time period.