[NNTP] Re: Nearly there

[Clive D.W. Feather]

> I have a vague memory that we had a couple of changes we were considering
> making at this stage, but can't remember exactly what they were. If anyone
> else can, this would be a good time to carry out a sanity check in the
> group.

Okay, trawling through the archives ...

(1) Keywords (which include capability names) are limited to 12 characters,
but we already have a keyword ("IMPLEMENTATION") which is longer than that.
The majority view was to remove the limit completely:

    A "keyword" MUST consist only of US-ASCII letters, digits, and the
    characters dot (".") and dash ("-"), and MUST begin with a letter.
    Keywords MUST be at least three characters.

      keyword = ALPHA 2*(ALPHA / DIGIT / "." / "-")

(2) What should the limit be on article numbers? At present it is 2^32-1,
but there were proposals:
- to make it 2^31-1, because some existing implementations use signed 32
  bit types to hold the article number and we should be documenting
  existing practice;
- to make it 2^64-1, either as a SHOULD or a MUST;
- to have no limit, but warn that existing implementations have limits.
There were also proposals for solving the limit (my BIGNUM proposal and
someone else's wrapping proposal), but these were not intended to go into
this document.

I'm very nervous about raising this, after all the arguments this time last
year, but I believe that changing it to 2^31-1 is the correct thing to do
*FOR THIS VERSION OF THE NNTP SPECIFICATION*. In support of this, I note
that Russ quoted our charter:

| The first concern of this working group shall be for the interoperability
| of the various NNTP implementations, and therefore for clear and explicit
| specification of the protocol. It is very important that we document the
| existing situation before taking up any new work.

and I also note something else he wrote earlier in the discussion:

| (I wonder how much software will fail at 31 bits, too.  I dug a lot of
| signed integers for article numbers out of INN when I started doing
| cleanups, and I think there are still a few left in the corners.)

(3) It was noted that some servers may be getting near this limit and we
don't yet have a good mechanism for dealing with this. It was therefore
proposed to add wording saying that this is unsatisfactory but is being
worked on ready for the next version.

    It is likely that at some stage at least one NNTP server will reach the
    limit of [4,294,967,295] in at least one newsgroup. Ways to deal with
    this while not gratuitously breaking compatibility are still being
    investigated and are likely to result in publication of a revision or
    extension to this specification at some future date.

    While servers MUST NOT send article numbers greater than this limit,
    client and server developers are advised to use internal structures
    and datatypes capable of handling larger values in anticipation of
    such a change.

-- 
Clive D.W. Feather  | Work:  <clive at demon.net>   | Tel:    +44 20 8495 6138
Internet Expert     | Home:  <clive at davros.org>  | Fax:    +44 870 051 9937
Demon Internet      | WWW: http://www.davros.org | Mobile: +44 7973 377646
THUS plc            |                            |