[Sigia-l] Core set of classification terms

[Jon Hanna]

> The above is an interesting point. There can definitely be problems with
> classifications translating. However, I feel that the Internet is
> a strong
> driver in the internationalization of language (the Englishization, to be
> more precise.)

So by "internationalisation" you actually mean the exact opposite of
internationalisation.
This does not bode well for theories of classifications scaling well.

> Words have often crossed over from language to language. (English
> borrowed
> a huge range of French words.) Indeed, the Irish for a person of
> black skin
> is "fear gorm."

Well lets not forget the mna gorm as well :)

Better still lets forget about language as I was using that as an analogy
and if we spend too long in analogy we will prove something in the analogy
that doesn't actually transfer to classification. While i18n is one case
where problems with classification do arise in the real world, they also
appear when the jargons of different fields occur, when different
technological identification systems arise, and so on.

In some cases one of these can be mapped pretty well with another, for
example the UUID URN would give a URI for everything with a UUID, and so
make everything with a UUID identifiable with a URI-based system such as RDF
(whether this is the best way to make that mapping is another matter).

In others we have much more difficult cases, and attempts to map one
classification to another will be problematic, often to the extent that it
is better to make no such attempt and instead allow both classifications to
exist side by side. Of course to allow side-by-side classifications we need
to decouple the resource from the classification, or at least decouple it
from all but one (there may be advantages in having one "real"
classification as long as that architecture can be made invisible to a user
of another classification).

> What I'm noticing is the emergence of a core set of web-based
> classifications, regardless of the language. Classifications like: Home,
> About, Contact, Feedback. Even if this is not always the case, it is
> certainly true that within English-language websites, these terms are
> beginning to dominate.

These are pretty broad, pretty simple, and very often separate from the
majority of the information (indeed I generally think of "Home" as having no
classification, but rather being an entry point into the classification(s)
on offer). These are generally part of the framework of a site, and to a
great extent belong more to implementation than content or knowledge.

I'm not saying you're wrong, but rather that the classification that
contains home, about and contact only go so far. They are important for
usability within the first 2 minutes of accessing a site, but beyond that
they don't really come into play.

Deeper into the system agreed on classifications like NewsML subjects, MESH,
etc. can be extremely useful to those familiar with them. They require one
to become expert to a certain extent, but if overlapped with other systems
then at least experts on a particular field will have a system that matches
their expertise without necessitating the same level of expertise in other
users. Of course they only go so far as well, MESH is probably a
prerequisite of a system dealing with the sort of medical information it
classifies, but little use even on a consumer medical site, never mind
beyond that.

> The Web often reminds me of airports. It's an international space; people
> are always moving from one website to another. Now, when you're
> designing a
> new airport, I doubt the design team sits around asking questions like:
> What are we going to call the Exit sign? What will we call
> Departures? What
> will we call Arrivals?

A lot of thought has gone into the signs in airports, and people still get
lost.