Moin! On 28 Nov 2025, at 17:00, Tobias Fiebig wrote: > I personally think that going for the option you suggested would open > another box, and likely be difficult for, e.g., Ralph in a similar > sense as the current version for you. The most both-sides-agreeable > option would, imho, be an acknowledgement and ref to RFC9539.
We should not quote experimental RFCs when we have standard track RFC that describe these transport mechanisms (RFC8484, RFC9250). And resolvers had used these before RFC9539 was out. In the end as these transports have the same problems with the underlying transport encapsulation (UDP/TCP) and solve them in similar ways, which may not something an software engineer has to care about when he uses a QUIC/HTTP library, but certainly is something to consider if you do your own DoH/DoQ library. Not that this is something one should do, but if you do there are drafts and documents out there how to deal with that on the TCP, HTTP or QUIC layer. It is not something that belongs into this document. So maybe saying something like other protocols that carry DNS payloads have their own mechanism to deal with these and are out of scope of this document. So long -Ralf ——— Ralf Weber _______________________________________________ DNSOP mailing list -- [email protected] To unsubscribe send an email to [email protected]
