Topic Maps are SGML or XML documents that describe what an information set is about, by formally declaring topics, and by linking the relevant parts of the information set to the appropriate topics. Since they can be in separate documents, and since they can work without changing an information set, we say that Topic Maps can be applied from "above" the information set, rather than from "inside" them; they are superimposed views. A topic map expresses someone's opinion about what the topics are, and which parts of an information set are relevant to which topics. There is no limit to the number of topic maps that can be created above the same information set.
Topic Maps help organize and retrieve online information in a way that can be mastered by information owners and information users. They play the same role as indexes play in books books, and that thesauri play in editorial consistency management. Topic Maps aim to dramatically enhance the efficiency with which human beings can find the information they need. Topic maps can support (i.e., can be formatted as) a wide variety of finding aids, including indexes and glossaries, both printed and online/interactive. The information sets for which topic maps can be finding aids can be large or small, including both technical manuals and enterprise-wide (or even civilization-wide) databases of documents.
Topic Maps are expressed in SGML or XML (the interchange syntax of Topic Maps is, in fact, an SGML and/or XML document type definition). An information set that is provided with a topic map can be expressed in any notation, including but not limited to XML and SGML. Some effects that are similar to the effects of topic mapping can be achieved by converting an information set into XML and inserting tags that identify and describe the topics to which each passage is relevant. This method is impractical and/or expensive for a variety of reasons; topic maps make this method obsolete and unnecessary. The information sets do not have to change in any way in order to be equipped with one or more topic maps.
Topic Maps are simply documents, and anybody can author one. The publisher of an information set can provide a topic map with the set; this will help the set's users find what they want, in much the same way as an index in a book. Users of an information set can also create their own topic maps, so as to organize their access to the information set in terms of their own desired topics, and their own desired connections among those topics.
Since topic maps can be merged, it is also possible that responsibility for a topic map can be delegated to several different people, each of whom produces a topic map without necessarily even knowing that it will be merged with other topic maps.
There are many small and large reasons for regarding Topic Maps as an important development. Here, we just mention some of the most economically significant ones.
They fulfill a universal need, and they are both powerful and simple. They lead the XML paradigm forward, opening the way for a "global positioning system for the Web", as Charles F. Goldfarb put it. They enable multiple alternative models of knowledge domains to coexist, and to work together, in a way that has not been available before. Topic Maps are capable of supporting and revealing immensely complex interrelationships within and among the concepts related to various fields of endeavor, and to provide master indexes to arbitrarily large and comprehensive bodies of information. The Topic Maps international standard (ISO/IEC 13250:1999) provides a way for information management system vendors to allow their finding aids to interoperate, and this breaks down barriers to information access between customers of different vendors. Topic Maps also heralds the day when there will be more and more comprehensive indexes of human knowledge, and when such indexes will be easier and easier to use, because they will be able to suppress irrelevant information more efficiently than ever before, and with a minimum of effort on the part of the user.
To subscribe to the topic map mailing list, send e-mail to Michel Biezunski, mb@infoloom.com
Last updated: September 7, 1999
FAQ maintained by Michel Biezunski and Steven R. Newcomb.