>1)  applications may query for "alt." for some RR. If the
>    resolver is not "alt-tld aware", or alt-tld functionality
>    has been disabled, then the request will return NXDOMAIN;
>    if instead the mechanism is supported, it would return
>    NOERROR with an empty answer. Note that this is backwards
>    compatible with non-alt-aware implementations since they
>    would leak in DNS and NXDOMAIN will be returned by some
>    other server.

OK.

>2b) as above, but it might be specified that the query is
>     to be done on a TXT RR specifically and that, instead
>     of the empty answer, a record might be returned with
>     contents that can be used to in some way ascertain that
>     the handler is actually for the implementation we expect
>     it to be (for example a software name or version, or
>     protocol "UUID"?)

I fear we are tumbling down the slippery slope into a tar pit here.

There is absolutely no standardization or commonality in the protocol
switches that people use for non-DNS names.  For .local it's in
library routines like gethostbyname() that map names to IP addresses,
for .onion it's in a SOCKS proxy that opens a connection to a remote
host and for those of us old enough to remember .uucp and .bitnet, it
was in a switch table in the mail configuration that determined what
program to call to do the next hop.  It's not much use providing an IP
address lookup handler for something that doesn't use IP addresses.

I hate to bring this up, but we might want to reconsider whether .alt
is still the best name for this hack.  For us old nerds, .alt reminds
us of usenet and uucp of yore, but in American politics it's some
unpleasant fringe groups with which I would not want to be associated
even a little bit.  Keeping in mind that we want to encourage people
to use .alt, I'd rather find something with no political (or other)
connotations.

I would suggest .app except that Google paid ICANN $25M for it last year.

R's,
John

_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to