At 00/12/04 08:15 -0500, Dave Crocker wrote:
Thank you. I was hoping someone would point out the support for parallel
operation so we could go further down that path. As you note, it seems to
be the closest to providing local/global support already.
That means postal gives us:
1. Global support for a common "character set"
2. Global support for a carefully mixed character set -- though really it
is just a partitioning between the global field and the local field
3. Local support for a local character set.
(the support goes beyond character set, but let's leave it at that if
that's ok.)
An immediate problem with comparing to postal is that it somewhat
correlates with the path a letter will take, so that the incremental
interpretation can be done by groups with different language skill-sets.
Really big post offices have special places to handle things such
as incomplete addresses. Nothing guaranteed, but if you are lucky,
you may even successfully send a letter from an arbitrary place to
anywhere in the world using local addressing, at least if you don't
forget the country name in the local script.
The DNS does not have that flexibility and the domain name interpretation
is not part of the transfer sequence of the data.
Yes, there are quite some differences. The advantage we have is
that as soon as the characters are somehow in the computer,
everything else is mechanical. This means there is no need
for a global field; if somebody is able to type in the address,
that's it, the machine does the rest.
Schemes that put an ACE-like field into a .com might be considered to be
like #2, above, by really they are not. The whole string is still global.
ACE is (maybe) for machines. It's not primarily intended for humans.
We may have ACE all the way (including TLD). It might be usable as a
poor man's ASCII equivalent, but I strongly doubt that anybody will
want to have it on the Latin side of their name card.
Frankly this leaves me viewing the postal example as pretty unhelpful for
finding a solution to the DNS requirement.
Well, the postal example shows how Latin and other scripts can
both be used to address something. The mixed case is not too
important for us, as discussed above.
In the postal example, conversion from one notation to the other
is a complex process (in particular for Japanese, lookup in context
is absolutely necessary). So I don't expect that something purely
mechanical (e.g. ACE) will do for DNS.
On the other hand, this thread was triggered by Graham's question about
the negative impact of partitioning. The postal example would seem to
show that the effect is not so bad.
Except I would claim that it is not partitioning. Note that an address
always has a global representation, in addition to a possibly different
local one.
It's a kind of partitioning, in that it is not always easy,
for everybody, to do use the 'local' address or to convert
from a local to a global one.
Perhaps that can reconciled as easily as claiming that any 'local' domain
name must also have a global form? (But, somehow, the word "scaling" gets
in the way of believing that.)
Scaling would be only by a factor 2.
Regards, Martin.