On Sunday 27 January 2008 21:02:33 Tony Hansen wrote:
But MIME sort of blurs the lines between the identity definitions
and the encodings.
Agreed to the extent that a single keyword either indicates an encoding
transformation which has an implicit domain, or the domain of unencoded
And you sort of blurred the lines by only talking
about MIME encoding and then using that to say that "extending the line
length limits" was incorrect.
No, I stated some objections to the implication that "implementations
which display messages" must "handle an arbitrarily large number of characters
in a line" (whereas MIME specifically limits lines using non-identity
encodings to 76 characters, specifically and as explicitly stated in the
MIME RFCs to provide a bound on what implementations must handle) and stated
that "extend this specification" might be misinterpreted in that context.
I did not use the term "incorrect" in that regard.
However, I don't think it's necessary to mention all potentially
relevant RFCs and errata. This particular paragraph was using the MIME
RFCs as examples.
Agreed that it's unnecessary to mention all potentially relevant RFCs.
However where normative references have errata or are amended by other
RFCs, those errata and/or amending RFCs should be mentioned, as they are
essential to correct interpretation of the (amended) normative references.
Speaking of the particular examples chosen, two of them are media type
registration procedure specifications, which are not relevant to the message
format specification per se.