ietf-nntp LIST OVERVIEW.FMT

[Charles Lindsey]

In <ylsmtmlcce.fsf at windlord.stanford.edu> Russ Allbery <rra at stanford.edu> writes:

>I personally think we should not standardize it because the contents of
>the database can change, which means that it's very unclear to me what
>utility the result has.  I think we have three basic options:

> * Standardize the command as written in our draft, or omitting the option
>   of presenting Lines and Bytes using the new metadata syntax, with the
>   warning that it's just what's going into the database now.

> * Standardize the command, making it optional and requiring that the
>   information be accurate for the entire overview database if it is
>   implemented.  But this still doesn't appear to be sufficient to allow
>   clients to draw conclusions like "Content-Type is in overview.fmt,
>   there's no Content-Type in the overview data for this article,
>   therefore the article doesn't have Content-Type" since the server
>   administrator may run some command that adds Content-Type globally to
>   the whole overview database between the time that the client sends
>   LIST OVERVIEW.FMT and the time that it retrieves that overview
>   information.

> * Don't standardize the command.

There is a fourth option:

  * LIST OVERVIEW.FMT tells you what the overview will contain for
articles added to the database NOW (which will likely be the same as for
all recent additions). You then give a warning that articles received in
the past MAY have used a different OVERVIEW.FMT. I would imagine this
pretty well accords with what current implementations actually do.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133   Web: http://www.cs.man.ac.uk/~chl
Email: chl at clw.cs.man.ac.uk      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5