[rfc-i] issue: canonical formats
Brian E Carpenter
brian.e.carpenter at gmail.com
Wed Jun 6 02:10:13 PDT 2012
On 2012-06-06 09:19, Yoav Nir wrote:
> On Jun 6, 2012, at 11:04 AM, SM wrote:
>> At 00:06 06-06-2012, Brian E Carpenter wrote:
>>> (off list)
[I changed my mind but forgot to remove that.]
>>> On 2012-06-05 12:26, Martin J. Dürst wrote:
>>> I don't see how we can "deal with that" except by stating which one
>>> is valid.
>>> And as Martin's story of two I-Ds and a patent seems to show,
>>> even formatting differences can be relevant. So the argument that
>>> automatic reflowing doesn't matter is unconvincing to me.
>>> I remain convinced that we need to state which version is canonical.
>>> That doesn't mean I'm against reflowable formats, etc., for
>>> convenience. It's just that there needs to be a reference version.
>> The RSE only has to agree about which version is
>> canonical. It can be a version which nobody
> I think it at least should be the one the author(s) review in AUTH48. That way, at least the canonical version has a higher chance of being error-free.
Exactly. It doesn't have to be the one for daily use, but it does
need to be the one that was formally "signed off".
>> It would be good if the content is as
>> similar as possible with the version people
>> prefer. It does not have to be an exact
>> copy. It does not have to be 40 characters per
>> line. It does not have to contain page numbers after every 2000 characters.
Right. But in case of technical ambiguity, it will be the signed-off
version that wins, so we'd better do the utmost to avoid ambiguity.
>> An alternative to all this would be "go and write
>> code". Natural selection will choose "the" version.
> Not going to happen. If there's one thing that these threads made clear, it's that Martin Rex and the people who read documents on iPhones not going to use the same output format.
>> There is this saying: "We reject kings,
>> presidents and voting. We believe in rough
>> consensus and running code". The first sentence
>> was about kicking out the IAB. The running code
>> part is whatever the RFC says, it does not matter
>> if the code does something else.
> Strange, I've never thought of equating "whatever the RFC says" with "running code". There's a bunch of thinly- and not-implemented RFCs out there, and a bunch of non-conformant implementations.
> rfc-interest mailing list
> rfc-interest at rfc-editor.org
More information about the rfc-interest