I was puzzled that articles on 303’s, such as Cool URIs for the Semantic Web never mention topic maps. Then I remembered, topic maps don’t need 303’s!
Topic maps distinguish between URIs used as identifiers and URIs which are the addresses of resources.
Even if the Internet is down, a topic map can distinguish between an identifier and the address of a resource.
Topic maps use the URIs identified as identifiers to determine if topics are representing the same subjects.
Even if the Internet is down, a topic map can continue to use those identifiers for comparison purposes.
Topic maps use the URIs identified as subject locators to determine if topics are representing the same resource as a subject.
Even if the Internet is down, a topic map can continue to use those subject locators for comparison purposes.
You know what they say: Once is coincidence, twice is happenstance, three times is an engineering principle.
The engineering principle? And its consequences? Keep watching this space, I want to massage it a bit before posting.
*****
Techies see: kill -9 ‘/dev/cat’ (Robert Barta, one of my co-editors).
Non-Techies/Techies see: Topic Maps Lab.
Spec readers: XTM (syntax), Topic Maps Data Model.
Not all there is so say about topic maps but you have to start somewhere.
*****
Apologies! News on CTM (Compact Topic Map syntax) most likely tomorrow. Apologies for the delay.
There is a dialectic tension between medium and message. There is no communication without both, and there can be no unambiguous communication in the absence of a distinction between them. The best way forward is to acknowledge and honor that dialectic tension, thus to make it exploitable. That’s the Topic Maps way, and it’s the subject-centric way. In Topic Maps, the message and the medium are distinct, and they are explicitly linked together in multiple, diverse universes of discourse, including medium-specific universes.
Topic Maps (ISO 13250:2000) make no assumptions about the medium of transmission. XTM and TMDM do make such an assumption (the medium is assumed to be the World Wide Web), but they are both application-specific; they *apply* a much broader, much more abstract paradigm.
Since its inception, the leadership of the W3C has sought to ignore the distinction between a medium-specific address (such as a URI, the medium here being the World Wide Web) and a semantic. This can be seen as a strategic behavior of an institution whose identity has far more to do with a specific medium, and those with interests in the development of that medium, than with the general idea of promoting human communication and understanding. Now things are changing, with the 303 business and “linked data”, but not in any fundamental way. TimBL’s first principle is still to use URIs to identify things that are not Web addresses. And that’s the problem, right there!
Apparently, the strategy has been to make human communication indistinguishable from the medium of the World Wide Web. Can anyone explain why that would be a good thing? Or why it wouldn’t be better to distinguish, always and everywhere, between medium and message?
I can explain why it’s a bad idea to confound URIs and semantics: it’s bad because it Balkanizes the representation of semantics. Every URI is owned by some specific interest. The WWW, as a medium whose doctrines are defined by W3C Recommendations, offers no way for the public at large to own the representation of semantics.
That situation must change, and it will change, one way or another. We can do better, and we don’t have to replace the WWW to do better. We merely have to recognize that the WWW is a medium, and that the medium is not the message. We have to be willing to avoid and decry all confusion between system addresses (URIs) and semantics. Whenever we do that, we strike a blow for the public good. (And, unfortunately, we annoy many vested interests. But that’s to be expected.)
Let’s resist the selfish temptation to invent ever more private languages, and instead work to overcome the barriers to human progress that linguistic diversity creates. Let’s embrace diversity, and honor it, and exploit it, but let’s not believe those who claim that a medium that offers no alternative to creating ever more linguistic diversity will somehow fundamentally bridge the cultural gaps that bedevil humanity. Experience shows that Esperanto is, after all, a fantasy. An Esperanto in which each word is owned by some interested party is a nightmarish, quasi-Orwellian fantasy, but it, too, is still just a fantasy. Let’s do things that really aid human understanding and that create and preserve public understandings and ideas.
(With due apologies to McLuhanites!)
Steve Newcomb
March 30, 2010
Comment by Steve Newcomb — March 30, 2010 @ 10:40 am
I don’t disagree with anything that you have said but am not sure that moral exhortation is the best approach.
I am reminded of a story from Cross and the Switchblade by David Wilkerson, which has a parable about religious conversion that is applicable here.
Imagine that you see a dog with an old, dry bone. The bone isn’t good for the dog but if you try to take it away, the dog is likely to bite you. Now, imagine the same situation but you drop a steak down next to the dog. The dog will drop the bone at risk to you and pick up the steak.
Rather than trying to take the bone of confusing semantics and system addresses, let’s use topic maps as the steak that causes people to drop that bone on their own accord.
Users who do will gain advantages over those that don’t and so that problem will solve itself.
Comment by Patrick Durusau — March 31, 2010 @ 9:17 am
[…] No 303’s for Topic Maps? I mentioned that distinguishing between identifiers and addresses with 303’s has […]
Pingback by One Billion Points of Failure! « Another Word For It — March 31, 2010 @ 9:18 pm