Re: ietf-nntp Three proposals

[Ofer Inbar <[e-mail]>]

[e-mail] (Ben Polk) writes:
> I think this is what the "Remove the X" group is proposing:
> 
> Today:     XOVER defacto manditory
> Tomorrow:  OVER manditory, XOVER manditory but depricated
> Day after: OVER manditory, XOVER optional but depricated
> Future:    OVER manditory, XOVER prohibited

I see it a little differently:

1. Today
  XOVER not documented in the standard, but de-facto mandatory for
servers and optional for clients.

2. Next year
  XOVER documented in the standard, mandatory for servers but
deprecated for clients.  OVER introduced, mandatory for servers.
XOVER is just an alias for OVER.

3. A few years later
  XOVER alias no longer used by many clients, becomes de-facto
optional for servers (but still widely implemented).  OVER used in
most clients and servers.

4. Sometime after that
  A new version of the standard document deprecates XOVER for servers
and prohibits it for clients.

So what we're proposing now is the step from 1 to 2, and no breakage
is possible there.  The breakage would come in stepping over from 2 to
4, if we don't wait for 3 to happen first.  If 3 never happens, well
then we can't ever move to 4.  But I think 3 will happen if we wait a
few years for it.

Besides, how many newsreading clients are there out there which can't
deal with the absence of the current X commands (XHDR, XOVER)?  Those
commands aren't in the standard, so clients which have no fallback to
deal with the absence of those commands are violating the standard.

Can we come up with a good list of these clients and how they fail?

Hmm, this .sig is strangely appropriate :)

  --  Cos (Ofer Inbar)  --  [e-mail]  [e-mail]
  --  WBRS (100.1 FM)   --  [e-mail] http://www.wbrs.org/
  "If I never had existed, would you still remember me?"
  "Would you think I was a burden on the state economy?"
     -- Austin Lounge Lizards, "1984 Blues"