open ends in blip 0032

[AndySchroder]

blip-0032 says

blips/blip-0032.md

Line 86 in b6c3e8c

Finally, a forthcoming bLIP will define a method to include the `user` part of the human-readable

and

blips/blip-0032.md

Line 118 in b6c3e8c

The node receiving the `invoice_request` can use the `user` field to determine for which user the

Where is this blip? Are these just added to the BOLT12 messages?

Also stated is

blips/blip-0032.md

Line 117 in b6c3e8c

a wildcard record may be provisioned as *.user._bitcoin-payment.`domain` with the same contents.

Why is this not defined in https://github.com/bitcoin/bips/blob/master/bip-0353.mediawiki ? Also what is the resolution order of the wildcard vs non-wildcard record? Seems like this needs to be defined.

Also, I think there should be some guidance on offer expiry and DNS record TTL. The record TTL should always expire before the offer.

Generally, I think the connection between BIP-0353 and blip-0032 is a bit loose and the should be referenced more tightly. Also, it might make more sense to separate the payer protocol from the DNS proof records over onion messages in blip-0032 into seperate blips (or at least make the blip title more inclusive of both topics). It is a bit confusing to have these mashed together and it might be easier to connect to BIP-353 if there is more clear separation of topics.

[TheBlueMatt]

I ended up proposing the name inclusion in BOLTs as lightning/bolts#1180.

...wildcard...
Why is this not defined in https://github.com/bitcoin/bips/blob/master/bip-0353.mediawiki ?

Its not super useful outside of the lightning context (otherwise all "user"s will end up paying to the same address with no way to differentiate).

Also what is the resolution order of the wildcard vs non-wildcard record? Seems like this needs to be defined.

DNS specific records override wildcard records (and in fact if you receive a wildcard DNSSEC proof you must validate that the proof contains non-existence proofs for the specific record).

Also, I think there should be some guidance on offer expiry and DNS record TTL. The record TTL should always expire before the offer.

Good point.

Generally, I think the connection between BIP-0353 and blip-0032 is a bit loose and the should be referenced more tightly.

Do you have any specific recommendations? Not sure what else to add.

Also, it might make more sense to separate the payer protocol from the DNS proof records over onion messages in blip-0032 into seperate blips (or at least make the blip title more inclusive of both topics).

bLIP 32 doesn't define the payer protocol at all, it only describes it in the non-normative discussion sections for completeness. It seems pretty separated given its in two totally separate sections of the document?

[TheBlueMatt]

Opened bitcoin/bips#1672 and #48