request shorter frame code points before RFC publication

[marten-seemann]

A 2-byte code point would be nice.

[huitema]

I do not really understand the process either. I was expecting some kind of form, but apparently we just need to send an email to the IANA, explaining that we need a code for a transport type and also two consecutive codes for the frame types. If we have to propose values, I would go with 0xAD1 for the "Address Discovery" TP, and 0x0B5A and 0x0B5B for the observed addresses V4 and V6 (OBServed A and OBServed B), but that's just because I like mnemonics.

But I wonder whether we should wait for WG adoption... which my take a bit more time.

[marten-seemann]

I agree. We can always change code points and bump the draft version.

[marten-seemann]

@nibanks had a suggestion for a particular 2-byte code points. Would you mind posting them here?

[nibanks]

I started implementing this in msquic and I just shortened what you had to:

    QUIC_FRAME_OBSERVED_ADDRESS_V4  = 0x9f81ULL,
    QUIC_FRAME_OBSERVED_ADDRESS_V6  = 0x9f82ULL,

[huitema]

@nibanks these numbers are larger than 2^14, and will be encoded on 4 bytes. If we want two bytes, we should use something like:

    QUIC_FRAME_OBSERVED_ADDRESS_V4  = 0x1f81ULL,
    QUIC_FRAME_OBSERVED_ADDRESS_V6  = 0x1f82ULL,

[nibanks]

Ah, you're right. I'm fine with those values you suggest.