[DNSOP] Re: I-D Action: draft-ietf-dnsop-domain-verification-techniques-09.txt

[Shumon Huque <[e-mail]>]

On Mon, Jul 21, 2025 at 3:33 PM Joe Abley <[e-mail]> wrote:

> Hi Shumon,
>
> On 21 Jul 2025, at 14:28, Shumon Huque <[e-mail]> wrote:
>
> > Yes, it means that you should concatenate multiple RDATA strings within
> the same TXT record.
> > We'll fix the wording here shortly.
>
> If we retreat briefly to the warm comfort of RFC 1035, a single resource
> record includes a single field called RDATA, so "multiple RDATA strings
> within the same TXT record" is a strange phrase.
>

I was partly suspecting someone might nitpick my quick email (I was going
to be more precise when composing the actual PR). You did not disappoint
Joe :)

How about changing:

"If there are multiple RDATA strings for a record, the Application Service
Provider MUST treat them as a concatenated string."

to:

"If there are multiple character strings in the TXT record data, the
Application Service Provider MUST concatenate the strings before use."

Here is another example from the DKIM RFC (6376):

"Strings in a TXT RR MUST be concatenated together before use with no
intervening whitespace."

RFC 1035 section 3.3.14 defines a field TXT-DATA as "one or more
> <character-string>s". <character-string> is "a single length octet followed
> by that number of characters". Also "<character-string> is treated as
> binary information, and can be up to 256 characters in length (including
> the length octet)".
>

I don't think we need to go into the details of the wire format of TXT
record strings containing a leading length octet. I think reasonable people
can understand that string concatenation will not include those length
octets.

The draft currently contains many examples of the word "RDATA". I haven't
> looked closely at every occurance in context, but I suspect it might be a
> good idea to define a new phrase like "domain validation string" to use in
> most or all of those casses, and to define it carefully,


That is worth considering. I'm  also tempted to avoid using a term (RDATA)
that mostly only DNS geeks are familiar with. Maybe we can just say "record
data".

Shumon.