What registries might do (was: Consensus Call on Latin Sharp S and Greek Final Sigma)

Andrew Sullivan ajs at shinkuro.com
Tue Dec 1 19:50:08 CET 2009

On Tue, Dec 01, 2009 at 06:21:38PM +0000, Shawn Steele wrote:
> Technical gobbledegook.  If an end users types Microsoft.com, it
> gets them to microsoft.com, they aren't going to make a subtle
> technical distinction.  Most users can't even figure out that
> paypal.safe.com isn't really "safe", how do we expect them to
> understand subtle technical differences in casing behavior?
> Certainly the end-users expect case insensitivity, that's what we
> appear to have given them.

I don't expect end users to make such a distinction.  But I do expect
participants in the technical standards community to make such
distinctions, because they actually make a difference.  In most cases,
for instance, if you enter Example.com into your attempt to reach a
target site, the resolution request goes out on the wire as
Example.com.  Some resolvers will use a compression pointer back to
the Example label in the question section, so no matter what is in the
zone, you'll see Example.com in the answer.  Others use compression
pointers only to authoritative data, so you'll see the label as it is
spelled in the first entry in the answer you get.  Even more amusing,
that could change over time, in case someone has put (for example)
three NS records with different RNAME capitalization in the zone
file.  I have seen this.

We've given them rough case-insensitivity, and that's what they've
come to expect.  I doubt very much anyone ever expected case
insensitivity before they started to use these tools: it's contrary to
many other conventions in English.

> So am I understanding correctly: MICROSOFT.com will get me to
> microsoft.com if I happen to use a capital letters, but BÜCHER.de
> won't get me to bücher.de?  How is that supposed to make sense to
> end users?

I thought part of the point of the mapping discussion was precisely
that clients were supposed to do "sane things" in their environment.
You seem to be arguing that clients are not in a position to do that.
I suspect what will actually happen -- what has mostly happened
already, indeed -- is that everyone will just use lowercase all the
time.  People have come to notice the caps lock key for the purposes
of their passwords.  Why won't they come to know these issues?


Andrew Sullivan
ajs at shinkuro.com
Shinkuro, Inc.

More information about the Idna-update mailing list