pre-draft for language

This is a draft for the filtering language.  I managed to fill in a few
examples, but there are still a lot of consistancy errors that arose from
making a lot of changes in the grammar.  I'm posting it anyway because I
believe it'll be better to actually let people see it.

If you find obvious errors, let me know.  I'm SURE there are places that are
not internally consistant.  I'll work on it more tomorrow.

Thanks ...

-- 
                                           Tim Showalter tjs <at> andrew.cmu.edu

Network Working Group                                       T. Showalter
Internet Draft                                           Carnegie Mellon
Document: draft-showalter-sieve-XX.txt                        March 1997
Expire sometime after it's released

                    SIEVE: A Mail Filtering Language

Status of this Memo

   This draft has no standing and is for limited distribution.  Please

sieve draft

This is the latest draft of the SIEVE filtering language.  I've also dropped
a copy at http://www.club.cc.cmu.edu/~tjs/draft-showalter-sieve-00.txt.  I'm
in the process of submitting a copy as an internet-draft.

I plan to do some heavy editoral changes, but I think it's close to complete
in terms of content and features.  Rob has been bugging me to make a good
clean distinction between keywords and commands, and that should happen
before this becomes an RFC.

Any comments would be appreciated!

--

-- 
                                           Tim Showalter tjs <at> andrew.cmu.edu

Network Working Group                                       T. Showalter
Internet Draft                                           Carnegie Mellon
Document: draft-showalter-sieve-00.txt                        April 1997
Expire sometime after it's released

                    SIEVE: A Mail Filtering Language

Status of this Memo

Summary of Filtering BOF

Here is my recollection of what happened at the Filtering portion of the
Submit & Filtering BOF at IETF Memphis.  This is reconstructed from memory.

FILTERING:

Discussion of Tim's draft (posted to list and copies passed around).
Consensus: good start, make minor fixups and submit as I-D.  Discussion of
extension mechanism.  Consensus: define capability string, which each
extension adds to.  How client gets string not defined in this spec.  Could
be ACAP, could be shared file, could be magic.  Question: how to write
script such that it can work on different servers, with or without
extensions.  Consensus: need dynamic test for extension supported, either
REQUIRES or IF-EXTENSION or both.  Discussion on how script gets from
client to server.  Consensus: leave for a different draft.  Could be ACAP,
could be file (.filters), could be anything.

Discussion of where filtering occurs.  Consensus: during "final delivery",
which will differ depending on implementation, but is before message goes
into mailbox.

Discussion on specific elements of draft:  IF/ENDIF vs. "{" and "}".  No
resolution; arbitrary choice.  Discussion of blocks: needed or not?  No
consensus for them.  Discussion of BOUNCE action: do we need it?  Does it
generate DSN or MSN or neither?  Does it need parameter for text to be
returned?  Call for dropping BOUNCE since we have REPLY and TRASH.
Discussion on how to specify text of REPLY.  Series of lines with
continuation character?  Length-prefixed series of bytes?  Series of lines
ending in dot by itself (standard SMTP etc.)?  Discussion of line-endings.
CR vs. LF vs. CRLF.  Consensus: add ABNF to allow any of three.  Message
text is series of lines, ending in dot on line by itself.  Question:

Note: Filtering BOF was *Unofficial*

Sorry, I forgot to clarify that the "Filtering BOF" was unofficial.

Fwd: I-D ACTION:draft-showalter-sieve-00.txt

The sieve draft I posted (except for some really trivial changes) has gone
out as an Internet-Draft.

-- 
                                           Tim Showalter tjs <at> andrew.cmu.edu

---------- Forwarded message begins here ----------

Mime-Version: 1.0
Content-Type: Multipart/Mixed; Boundary="NextPart"
To: IETF-Announce <at> ietf.org
Sender: ietf-announce-request <at> ietf.org
From: Internet-Drafts <at> ietf.org
Reply-to: Internet-Drafts <at> ietf.org
Subject: I-D ACTION:draft-showalter-sieve-00.txt
Date: Mon, 21 Apr 1997 09:42:37 -0400
X-Orig-Sender: cclark <at> ietf.org
Message-ID:  <9704210942.aa04502 <at> ietf.org>

--NextPart

 A New Internet-Draft is available from the on-line Internet-Drafts 
 directories.                                                              

       Title     : SIEVE: A Mail Filtering Language                        
       Author(s) : T. Showalter
       Filename  : draft-showalter-sieve-00.txt
       Pages     : 19
       Date      : 04/18/1997

Sieve Readability Suggestions

I finally made time to read the sieve document carefully.  In general, I
think the language is in good shape.  I've got a few suggestions which I
think may improve sieve's human readability without changing the
complexity of parsing much. It's not bad as it is, but these might make it
slightly better.

(1) Comments

I suggest using "--" instead of "#" as the comment introducer.  "--" is
what Applescript uses, and is more likely to make sense to the average
human (rather than a programmer) IMHO.

(2) The "header" test

I'd like it if the <header-name-list> and the <key-list> items could be
either a list or a single quoted string.  Thus you would get:

if header "Subject" contains "$$" then
    toss
endif

or

if header ("Subject" "From") contains "B1FF" then
    toss
endif

(3) The "header" -nocase arguments

I think nocase should be the default with -case as the modifier.  So you'd

Re: Sieve Readability Suggestions

On Fri, 25 Apr 1997, Chris Newman wrote:

> (1) Comments
> 
> I suggest using "--" instead of "#" as the comment introducer.  "--" is
> what Applescript uses, and is more likely to make sense to the average
> human (rather than a programmer) IMHO.

I can see this, but I'd like a second opinion.  It seems a little strange to
me.

> (2) The "header" test
> 
> I'd like it if the <header-name-list> and the <key-list> items could be
> either a list or a single quoted string.  Thus you would get:
> 
> if header "Subject" contains "$$" then
>     toss
> endif

This is probably a good idea.

> (3) The "header" -nocase arguments
> 
> I think nocase should be the default with -case as the modifier.  So you'd
> get "contains", "is", "matches", "contains-case", "is-case",
> "matches-case".  This makes it more readable, IMHO, and makes the
> more common case easier to type.

Ok, this makes sense.

Re: Sieve Readability Suggestions

On Fri, 25 Apr 1997, Tim Showalter wrote:

> On Fri, 25 Apr 1997, Chris Newman wrote:
> 
> > (1) Comments
> > 
> > I suggest using "--" instead of "#" as the comment introducer.  "--" is
> > what Applescript uses, and is more likely to make sense to the average
> > human (rather than a programmer) IMHO.
> 
> I can see this, but I'd like a second opinion.  It seems a little strange to
> me.

Both are used in other languages, so there is a precedent for either
one.  I can see the argument for matching AppleScript, since there are
some otherwise novice potential users already familiar with it.  Do
any other languages targeted to novices use "#"?

> > (4) Trailing ";"
> > 
> > The grammar seems to require a trailing ";" after every command
> > (including an endif).  I'm starting to think it would be better to simply
> > have CRLF terminate a command as many of your examples do (and use "\"
> > CRLF to fold a line).
> 
> I'll fix the examples to reflect what's in the draft.  I'm not sure I like
> this change, but it might clean things up.  I'd like a second opinion.
> 
> I would drop ; from the grammar entirely if this is the case, i.e., no more
> than one command on a line.

Re: Sieve Readability Suggestions

David L Miller <dlm <at> cac.washington.edu> writes:
>On Fri, 25 Apr 1997, Tim Showalter wrote:
>> On Fri, 25 Apr 1997, Chris Newman wrote:
>> 
>> > (1) Comments
>> > 
>> > I suggest using "--" instead of "#" as the comment introducer.  "--" is
>> > what Applescript uses, and is more likely to make sense to the average
>> > human (rather than a programmer) IMHO.

Only to those familier with applescript.

>> I can see this, but I'd like a second opinion.  It seems a little strange
>> to me.

>Both are used in other languages, so there is a precedent for either
>one.  I can see the argument for matching AppleScript, since there are
>some otherwise novice potential users already familiar with it.  Do
>any other languages targeted to novices use "#"?

I'm not sure worrying about the familiarity of '#' to novices is a big
deal.  It's short and to the point and it's in wide use on at least one
platform (.  For users that are 'pathologically-novice' a GUI interface
is more in order anyway.
--

-- 
thx,
kjj

some input on sieve-00

0.2 - Open Issues.

I'm curious why "fileinto" is being considered for moving into a separate
document.  Maybe I missed an earlier thread on this.

1. Introduction

The word 'sort' in 4th paragraph is maybe a bit ambiguous when talking
about sorting mailboxes.  Maybe something like 'autofile' or something.
Not a big point, just a comment.

2.5.1 Headers
2.5.2 Addresses

There are notes that asks whether these sections is necessary or useful.
I think they are.

2.7 Evaluation

The second paragraph mentions the possibility that implementations may
impose restrictions on the number of actions per message.

I think this is a bad thing.  While I understand that one of the design
goals is to reduce the possibility of using the mechanism for mail-bombing,
I want to point out that there is sufficient existing practice to allow
such actions as auto-filing in conjunction with auto-forwarding.

With a restriction like this, the sieve becomes significantly less useful
for many of the users that are most likely to use it the first place -
users adept at email.