ietf-mta-filters
[Top] [All Lists]

Re: AD Evaluation Comments: draft-ietf-sieve-vacation-04

2005-10-20 11:19:55

Some questions and comments, also noted in the I-D Tracker.  Please clue me
in as appropriate:

Minor IDNit errors are noted in the tracker.  The shepherd write-up said
that IDNits ran cleanly.  Hmm...

The two nits appear to be additional spaces between words. I've removed them.
(For the record, I think this is a stupid and unnecessary check for idnits to
make, especially since the vagarities of line breaking can create situations
where extra space lurks indefinitely in a document.)

General comment: both "Sieve" and "sieve" are used in the document.  One
form or the other should be used consistently.

It's a proper name for a language so "Sieve" would appear to be the better
choice (except in file names, of course). Fixed.

First page:
The "??" characters used for Tim's organization need to be replaced.  To be
honest I'm not sure if anything should be used at all if there is no know
organization.

xml2rfc (incorrectly IMO - since we all participate as individuals, why can't I
be an entirely unaffiliated draft author?) insists on having an organization
field. I fired off a note to Tim asking him what he wants to have there. Best
I can do.

Section 4.1:
I'm curious about the "Sites MAY also define a maximum days value, which
MUST be greater than 7, and SHOULD be greater than 30" text.  Why use a MUST
for what appears to be an operational matter?  What is the impact on
interoperability if a lesser value is used?

Simple: In order to write portable scripts there needs to be some guarantee as
to how big a :days value you can specify. This requirement means a script that
specifies a value of 7 or less is guaranteed to run on any compliant Sieve
system that supports vacation. Without this requirement you don't have such a
guarantee.

Section 4.2:  "Implementations are free to limit the number of remembered
responses, provided the limit is no less than 1000."
Why 1000, and is this a normative requirement or not?

1000 seemed like a reasonable value. It is no more justified (or justifiable)
than, say, the 1000 character line length limit in SMTP.

I'll make the language normative.

Section 4.7: "A script will fail if it attempts to execute two or more
vacation actions."
Should "will" be either SHOULD or MUST?  "will" describes an implementation
behavior.

I'll make it a MUST.

                                Ned