Two further comments on this rsync example:
1) As written, and to conform to our general notion of stable reference,
if we want a normative reference to rsync, it would have to be a
reference to a specific version, with a reliable way that folks could
get to the definition of that version.
2) As written, and something I think is important, it would take a
further exception to treat the code as a specification. Code is NOT a
specification for interoperable implementation. And I would consider it
a serious flaw if the only acceptable implementation came from a single
source, even if that source was an open one.
Thus, if the rsync open source project included a clear and detailed
description of the protocol, then this change would allow that to be a
normative reference.
Yours,
Joel
On 10/30/17 11:53 AM, Eliot Lear wrote:
Hi Russ,
On 10/30/17 2:19 PM, Russ Housley wrote:
Alia:
The SIDR WG wanted to reference rsync, but there was no stable reference that described the protocol, only the code. A document was written to specify the rsync URL, and they is the only thing that is normatively referenced. A reference to the protocol would be better, and I hope this BCP would allow that to happen in the future.
When you say, “the protocol”, to what are you referring?
Section 2 is written as a list of questions. In some cases it is unclear to me which answer to the question makes an acceptable normative reference. A list of requirements would be more clear.
I think that's a fair point, but for a -00 the questions seem like a
good place to start. Maybe you even have some ideas as to what good
answers should be.
Eliot