Re: draft-iab-dns-choices-02.txt comments: host names vs domain names

[wayne <wayne@schlitt.net>]

In <06C10095E8598BB12246E048@[192.168.100.25]> Alex Bligh <alex@alex.org.uk> writes:

> --On 13 June 2005 15:48 -0500 wayne <wayne@schlitt.net> wrote:
>
>> The problem isn't as much in the applications as in things like DNS
>> hosters.  For example, GoDaddy (which recently passed NSI to become
>> the largest registrar) does not allow me to create a TXT record with
>> the name of _underscore_test.unified-spf.org.  The inclusion of the
>> underscore causes the webform to reject it.  They also don't allow the
>> creation of SRV records, let alone unknown types.
>
> Isn't this argument a bit backwards? IE don't these DNS hosting services
> prohibit use of underscores precisely because there is no standards-based
> protocol that mandates their use (and banning them for A records etc.
> eliminates a lot of tech support calls from the clue deprived). Neither do
> they support RR Types which have no standards-based definition.
>
> If you design the protocol right, vendors (software and hardware) will
> support it.

Ok, yes, you have a point.  The problem with DNS hosting services only
supporting hostnames and not domainnames will be solved if there is
ever a compelling enough reason for them to accept the customer
support headaches and customer education that they will have to do.

Once DNS hosting services update their software, future _prefix
schemes won't have to blaze that particular path.  They will still
have to educate the general public though about their protocol's usage
of _prefix.


> Case in point: most of these DNS vendors previously did not support SRV &
> NAPTR records. I am betting, now these are becoming more useful, you will
> see a roll-out (clearly they are already in Bind, and will over time get
> into hosted DNS products). Should these have been implemented by TXT too,
> because n years ago no vendors supported them?

I think that the deployment of both the SRV and NAPTR records would
have happened faster if they were based on the TXT## RR clone idea.
Last year I think IETF-59, the subject of SRV record usage came up in
conjunction with MARID.  The Jabber folks were queried about their use
of SRV records and they reported that there are still lots of
deployment problems with them.  SRV records may be used a lot in
intranets because a single company/organization controls all the
parts, but for an internet-wide protocol like Jabber is still running
into problems.


Of course, I we could probably easily start a flame war just by asking
"should NAPTR records have been done?" instead of why kind of records
they should use.


-wayne



--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>