Re: [lemonade] Comments on draft-ietf-lemonade-catenate-05.txt

Randall Gellens <randy@qualcomm.com> writes:
>In Section 4 (Response Codes), do we really mean "may return a 
>response code"?  I think "SHOULD return a response code" would be 
>better, but in any case it is not clear if "may" is intended to be 
>normative.

The catenate draft does reference RFC 2119 or otherwise define
particular words in uppercase as having special normative meaning.
Changing that sentence to use "SHOULD" (or "MAY") would require many
other changes to the draft to keep it consistent.  On reflection, I
agree with the author's choice to use normal English prose and let the
words speak for themselves.  No, really, we can use the words "may",
"should", and "must" as other than normative baseball bats to beat
implementors with.

With regards to that particular sentence, I read it as being
descriptive, acting as an introduction to the subsections that follow.

>In Section 4.1 (BADURL Response), it is not clear if "invalid" means 
>syntactically or semantically regarding UIDs and body parts.  That 
>is, is this response code to be used if the syntax is bad or if the 
>referenced item does not exist?

Given the first sentence of the subsection, it is obviously meant to
apply to *all* cases where a supplied URL cannot be resolved to a chunk
of text to include.  Perhaps the second sentence should be changed to
start "Some possible reasons..." to make it clear that the list is _not_
exhaustive.

Philip Guenther

_______________________________________________
lemonade mailing list
lemonade@ietf.org
https://www1.ietf.org/mailman/listinfo/lemonade