Stop me if I've misunderstood...

Andrew Sullivan ajs at shinkuro.com
Fri Jul 10 06:41:47 CEST 2009


On Thu, Jul 09, 2009 at 11:53:56PM +0000, Shawn Steele wrote:

> implement form.  (And doesn't DNS recommend lowercase now anyway?)

I don't know what "recommend lowercase" means in this context, but if
you mean, "Was there any BCP published that said 'use lowercase'," my
answer is, "Not that I know of."  I am prepared to be wrong (at my age
with my record, I sorta have to be!), but DNSEXT has been asked to
adopt a document that explicitly uses the 0x20 bit as a kind of
side-channel from a resolver back to itself.  And as I said, the data
that is actually in the publisher's zone may not make any difference
for how things display.

> > This is because most name servers use the compression trick
> 
> Sorry, I don't know what the compression trick is :)

In DNS, we have compression by labels.  The first time (for instance)
.com appears in a DNS message, it appears as all its
octets. Subsequent instances of that label are just references to the
first instance, which means you get the data however it was spelled
that time.  There are various rules about when compression may be
used, and new RRTYPEs aren't allowed to use it; but there's reason to
believe that it is the mechanism underlying many popular resolvers'
actual behaviour.

A

-- 
Andrew Sullivan
ajs at shinkuro.com
Shinkuro, Inc.


More information about the Idna-update mailing list