Open mnot opened 3 years ago
Given that most documents come from the IETF stream, it seems reasonable for IETF
to be the default, as it was in 7991 and 7749 before that.
I think seriesInfo can be used with documents that are not RFCs, and there can be several seriesInfo in the same reference, so this may not be backward compatible but it probably is better not to default it.
<seriesInfo name="BCP" value="14"/>
<seriesInfo name="RFC" value="2119"/>
<seriesInfo name="DOI" value="10.17487/RFC2119"/>
@jrlevine That's an interesting point about the possibility of including several seriesInfo elements in a reference. However, the same document can't come from multiple streams (AFAIK), so that seems orthogonal to the current issue.
As to the fact that "seriesInfo can be used with documents that are not RFCs", that might argue for not specifying a default value and also for not limiting the value only to members of the set specified in RFC 7991.
I think we should try define things that go into \
If we need stream information in references, we should try to understand the purpose, and how it's rendered, and then define it in terms that it works for other document sources as well.
AFAIU, we're really looking only for a way to insert some information in a specific part of the reference, right?
If stream has a default, then I believe this is wrong:
<seriesInfo name="BCP" value="123" stream="IAB"/>
<seriesInfo name="RFC" value="4567"/>
because the second stream defaults to IETF, which is a conflict. I don't think there is a way to say use a default unless some other element nearby has a value that overrides it. I agree with Julian that this would be better as a separate element to use where relevant. RPC said that they believe that readers find stream info in references to be useful.
I agree with both of you. :-)
Further discussion in the change management team leads to a suggestion that we remove @stream
from seriesInfo and update the documentation. Rationale: (1) we can’t have a single default for all the different series (2) RPC doesn’t use @stream
for the seriesInfo element (3) using a separate stream element is less broken than trying to mix the stream in with the seriesInfo element.
Given that most documents come from the IETF stream, it seems reasonable for
IETF
to be the default, as it was in 7991 and 7749 before that.
FWIW, there was no "stream" attribute in 7749.
(2) RPC doesn’t use @stream for the seriesInfo element
FWIW, that is not correct. It is used in seriesInfo for the document boilerplate (not inside \
In revision 2.32 of the RNC, the default value (of
IETF
) was removed fromseriesInfo@stream
. This is not backwards-compatible.