“Use My Model/Language Mister!” is the cry of markup, modeling and semantics projects.
They all equally sincere and if you don’t like any of them, wait another six months or so for additional choices.
I don’t remember if it was after the 75th or 100th or somewhere past the 100th “true” model that I began to suspect something was amiss.
Models and languages change over time and can be barriers to discussion and discovery of badly needed information.
Rather than arguing for this or that model, as though it were some final answer, why not ask which model suits our present purposes?
With topic maps, once the subjects under discussion are identified, how they are represented for some purpose is a detail. A very important detail but a detail none the less.
If, or rather when, our requirements change, the same subject can be represented in a different way. The subjects can be identified, again, to create a new representation, or, if identified using topic maps, our job of moving to another model just got a whole lot easier.
Hello, Patrick 🙂
I don’t really see any problems with there being many data models, a “shoe that fits all” model is probably impossible. To compare, there a loads of programming languages too, and the developers pick the ones that suit their paradigm. What I think is important is that all these models have a common manner of identifying subject. Another problem with pinning down subject definitions is On my blog I wrote a short post ( http://meronymy.blogspot.com/2010/03/linguistic-relativiry-and-subject.html ) about how a subject can give different cognative classifications, this is sadly something that a simple URI can’t prevent from happening and is a challenge for the future.
Comment by Inge Henriksen — April 28, 2010 @ 3:01 am
Hello Inge,
I’m not convinced that a “common manner for identifying subject.” is any more possible than having a common way to identify subjects. Your reference to linguistic relativity being a case in point.
Rather we need an expression of how subjects are identified, and we can map those expressions much in the same way we map subject identifications themselves.
Thanks for the pointer on linguistic relativity! Two quick impressions:
1) Simple URIs are at best a way to substitute for a collection of values that we think identify a subject, a shorthand if you will. (I rather like that, a shorthand for subject identity. May have to use that more often.)
2) To the extent that we “see” different colors, doesn’t that indicate that we are “seeing” different subjects?
I don’t find it troubling that we may have different views that don’t line up, at least in some views. Other mappings of the same views, may consider them to be identifying the same subject.
Accepting that subjects can be identified any number of ways opens up enormous archives of digital information not written using URIs to identify subjects.
I think of semantic diversity as something to be enjoyed, albeit that it makes the technical side more difficult, as opposed to being something to be overcome.
Think how boring it would be if we all wore white shirts and dark pants, everyday. Well, I do but I don’t advocate it for everyone. 😉
Comment by Patrick Durusau — April 28, 2010 @ 8:33 am