Re: whois++ response formats...

[Rickard Schoultz <[e-mail]>]

> Here are the response formats I am using for responses that
> don't involve templates -- culled from the other whois++
> implementations.  Let me know if you disagree strongly with
> any of this stuff :-)

Martin,
 I think your set of responses are as good as any.  My opinion is that
we need a BNF (Peter?) for the FULL response, and use that to form the
responses from system commands.

If we start out with:

# FULL <hits>
# <template> <handle>
 <attribute>:<value>
 ...
# <template> <handle>
 <attribute>:<value>
 ...
...
#END

and apply it to the different responses:

>   version
>   # FULL 1
>    Version: LUT version 0.0
>   # END

This would add another line, with a questionmark on what the handle
should be:

# FULL 1
# VERSION version
 Version: 1.0a
# END

>   commands
>   # FULL 16
>   # command describe
>    Name: describe
>    Description: Describe Server
>   # command help
>    Name: help
>    Description: Help!
>   [...]

... follows the model.

>   constraints
>   # FULL 13
>     search
>     case
>     format
>     maxhits
>   [...]

Here we have the opportunity to tell about how we support different
constraints, so I think it too should follow the above model like:

constraints
# FULL 3
# CONSTRAINT MAXHITS
 Constraint:     maxhits
 Default-Value:  200
 Range:          1..n
 Description:    Maximum hits returned.
# CONSTRAINT FORMAT
 Constraint:     format
 Range:          full,abridged,handle,summary
 Description:    Selects output format.
# CONSTRAINT HOLD
 Constraint:     hold
 Description:    Keeps the connection open.
# END

> Ciao,
>
> Martin

-Rickard