Re: [tm-pubsubj-comment] Baltimore minutes

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Lars Marius Garshol
|
| In fact, the two are fundamentally opposed, because having
| identifiers that are meaningful to humans means there will be
| reasons to change those identifiers based on what they mean to
| humans.

* John Cowan
| 
| This, however, goes too far.  As I pointed out a year or two ago,
| the human-meaningful identifier "Roma" has referred to exactly the
| same city for the last 2756 years, a record of stability unmatched
| by any meaningless identifier whatsoever.  AFAIK, certain
| house-numbers are the oldest such, and they date back a few
| centuries at most.

That's true, but I still think it is true that identifiers with
internal semantics carry within them potential instability in a way
that identifiers that lack this do not. As you point out, that does
not mean that identifiers without semantics are always more stable
that identifiers with semantics.
 
| What matters is that the authority which assigns the identifiers,
| whether meaningful or not, is or is not committed to stability. 

True. However, if that authority decides to also assign identifiers
with internal semantics it has created a problem for itself, as we've
just seen. There will be cases in which the semantics in some of the
identifiers will no longer be correct or desirable, whether because
the world has changed, or because our opinions about it have changed.

Let's say we assigned the identifier "it-roma" to Rome, to distinguish
it from other Romes around the world. If Italy at some point ceases to
exist, whether because it is split into smaller pieces (which Lega
Nord wants to do) or because it joins the EU, those semantics will
suddenly be inappropriate.

At this point the registration authority will have to either keep
"it-roma" (stability over semantics) or change to "eu-roma" (semantics
over stability). As you say, what matters is which of these two the
authority chooses, but choosing to honour stability over semantics
is also problematic.

Of course, there's also the issue of human-friendliness, and I think
ISO 639 and 3166 have made the right choices. Languages do not change
very often, so using human-understandable lables in ISO 639 is OK (and
would have been no problem had ISO chosen stability over semantics).
Countries, on the other hand, do change often, and even change names,
and so choosing numbers in ISO 3166 is also right.

| The numerical identifiers formerly used in the U.S. to identify
| kinds of businesses (now replaced by the NAICS codes) were regularly
| reassigned when new members were added to the set, due to a desire
| to have the human-readable labels come out in alphabetical order
| when the codes were sorted numerically!

I suppose you mean this as an illustration that if the authority is
not committed to stability having identifiers free of semantics does
not help. I'll grant you that.

To me it is beginning to seem that there are some general guidelines
that can be pointed towards, but that hard and absolute rules are
going to be very difficult to come up with.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >