Topics represent subjects (your concepts) and are a gathering point for information about that subject.
For example, "record linkage," "list washing," and "database hardening" all refer to the matching of records of different origins that are about the same subject. As an aside, record linkage originated in epidemiology, which needed to study mortality records across different record keeping techniques. Typically, the records are all converted to a common format/terminology for processing.
Contrast that with a topic map. The terms "record linkage," list washing," and "database hardening" could appear on a single topic, so that if you searched for any of those "labels/names" you could find all the information a topic map had on that topic. Such as a bibilography that has publications that use all those labels/names and others.
Distinguishing URIs between Identifier vs. Locators and modeling relationships are also strengths of topic maps.
I am the editor of the Topic Maps Reference Model (an ISO standard) but more accessible information can be found at:
Or you can contact me directly, Patrick Durusau, patrick@durusau.net (http://tm.durusau.net)
So the answer to your question is, that not enough developers know about them, or that topic maps are not useful enough to developers.
Can you give an example of how they can be used? How are your using them? Why are they useful to you?
I don't think it wouldn't be helpful to me to just organize my knowledge by topics/keywords. The tricky parts are the concepts, not the connections between topics/keywords.