Re: [apps-discuss] Gen-ART review of draft-ietf-appsawg-malformed-mail-09

["Murray S. Kucherawy" <[e-mail]>]

I think I'm going to request creation of my own private Directorate.  David
seems to get all my Gen-ART reviews, and we could possibly do some
optimizing here.  :-)

On Fri, Nov 1, 2013 at 6:32 PM, Black, David <[e-mail]> wrote:

>
> Nits/editorial comments:
>
> The word "naked" is used a few times to refer to something that occurs in
> isolation, without enclosure in another construct, e.g., a naked CR.  This
> idiomatic use of "naked" should be explained before it is used.
>

Fixed; just used "isolated" again.


>
> In Section 1.1, I have always heard Postel's law as:
>         - Be conservative in what you send, and
>         - Be liberal in what you accept.
> The change from "do" in this draft to "send" (above) seems useful, as
> it should help focus the discussion in the second paragraph of Section
> 1.1 - Postel's law, as I have understood it, has never blessed anything
> remotely resembling there being "no limits to the liberties that a
> sender might take."
>

There are a bunch of versions.  See:

http://en.wikipedia.org/wiki/Jon_Postel#Postel.27s_Law

Some of them say "do", some say "send".  Perhaps the most authoritative one
is here, in Section 1.1.2:

http://tools.ietf.org/html/rfc1122

It says "send".  I'll make that a reference and adjust.

I concur that the Law doesn't bless the extremes, but I believe the email
community has (deliberately or through ignorance) gone there, which is
largely the point of this work.


>
> Section 5's section title "Mail Submission Agents" doesn't seem to be
> connected to its MHS and MTA content.  It would be useful to add a
> sentence to remind readers, including this one ;-), of what Mail
> Submission Agents are.
>

The final sentence of that paragraph specifies what it does, but I see your
point about the disconnect in terminology.  Will adjust to tie it all
together.


>
> Sections 7.1.* offers degrees of advice qualified by "safely", "usually",
> "reasonably" and "should".  There appear to be only two concepts:
>         - "safely": Do this all the time.
>         - "usually", "reasonably", "should": This is the recommended course
>                 of action, but there may be exceptions.
> While RFC 2119 is not intended for Informational documents, this is an
> example of the sort of sloppiness that RFC 2119 is intended to clean up.
> At the very least, the use of three words for essentially the same concept
> is poor form, and RFC 2119 can be used in an Informational document when
> appropriate caveats are provided in the terminology section that references
> it.
>

Earlier versions of this draft looked roughly like an applicability
statement, and thus had RFC2119 language throughout.   We decided that, as
you point out, it's mostly advice, and not a way of establishing a
capability within Internet Mail, which would be more like what an
applicability statement is for.

I'm thus inclined not to backtrack, but merely satisfy your "At the very
least" clause.


> In Section 7.1.4, "Likewise" is not a good way to associate the second
> example with the first, because it handles the missing parenthesis in a
> rather different fashion (adds quotes instead of inserting the missing
> parenthesis character).
>

Removed.


>
> In Section 7.7, the first use of "8bit" occurs in "8bit material" but some
> of
> the subsequent occurrences omit the word "material" - that word should be
> used with all occurrences.
>

Fixed.


>
> idnits 2.13.00 generated a couple of warnings about obsolete references:
>
>   -- Obsolete informational reference (is this intentional?): RFC 1113
> (ref.
>      'PEM') (Obsoleted by RFC 1421)
>
>   -- Obsolete informational reference (is this intentional?): RFC  733
>      (Obsoleted by RFC 822)
>
> In both cases, the reference appears to be intentional, and the warning
> should be ignored.
>

Correct; I'll make sure the AD sees this.

Thanks!

-MSK