Re: [DNSOP] I-D Action: draft-ietf-dnsop-dns-capture-format-05.txt

[Stephane Bortzmeyer <[e-mail]>]

On Thu, Feb 22, 2018 at 12:31:29PM +0000,
 Sara Dickinson <[e-mail]> wrote 
 a message of 51 lines which said:

> We have an update to draft which we hope captures all the comments
> to-date.

I think so, too. At least, it captured mine :-)

> There are still a number of ‘QUESTIONS’ in the draft that we would
> appreciate feedback on.

>   QUESTION: Should fields be added to indicate the sampling/
>   anonymisation method used?  If so, it is proposed to use a text
>   string and RECOMMEND it contain a URI pointing to a resource
>   describing the method used.

Yes. We cannot use a registry because the number of possible methods
is high, and changing, and each method has parameters. So, the idea of
the URL seems OK.

>   QUESTION: Should there be another flag to indicate that names have
>   been normalised (e.g. converted to uniform case)?

Yes. Otherwise, you cannot survey if draft-vixie-dnsext-dns0x20 is
used ot not.

>  QUESTION: No EDNS(0) option currently includes a name, however if one
>  were to include a name and permit name compression then both these
>  mechanisms would fail.

Is it even possible? RFC 6891 does not mention it. But it says
"OPTION-DATA. MUST be treated as a bit field." which seems to imply
that storing the entire OPT RR, or the option data, as a blob, is
perfectly OK.

>  QUESTION: There has been no feedback to date requesting further work
>  on the processing partially malformed messages.  The editors are
>  inclined not to include it in this version.  It could be the subject
>  of a future extension.

I agree. We waited for this document long enough. 


Editorial:

> Adherence to the second two rules given in Section 3.9 of [RFC7049

Isn't it clearer to say "the last two rules"? "The second two rules"
(meaning the third and the fourth) confused me.