Re: [Fwd: Re: [Fwd: [db-wg] RIPE51 DB-WG Draft Agenda V1]]

[Denis Walker <]

Hi

The  irt object was implemented in a way to try to make it easy to use.
We did not want to have to reference an irt object from every inetnum
object.  So when an irt object is referenced from an inetnum object it
applies to all the more specific inetnum objects, until another irt
object reference is found.

With the -c flag we implemented a method of finding the related irt
object to any given range. But it does require some effort on the part
of the user to do this. You first query for the range to find out who
operates that range. Then you query again using the -c flag to find the
related irt object.

From this discussion it looks like what you want is for the first query
to return this irt object as the default, along with the inetnum object
that would normally be returned with this query.

If we make this the default operation, then we have to have a way to
turn it off. Not everyone wants yet more information returned with their
query. This could be done with another flag. Or we could link it in with
the recursive searching currently done with personal data.

This would mean a query done for a range with no query flags would
return the most specific inetnum object, route objects, organisation
object, role/person objects and the irt object that would have been
found with a -c query. The same query with the -r flag would only return
the inetnum and route objects.

Is this the behavior that most people would prefer?

regards
Denis
Software Engineering Department
RIPE NCC

Havard Eidnes wrote:

>>>We have changed the behaviour of the database so that when
>>>you do a "-c" lookup, we return the IRT object in the
>>>reply. I thought this was the desired behaviour. You can see
>>>it here:
>>>      
>>>
>>This is useful, but the desired behavious was to always return
>>the relevant irt record for inetnum queries, along with the
>>most specific inetnum object.
>>    
>>
>
>I guess the missing phrase both in the action point and in the
>above reply is "by default".
>
>(And, yes, I agree, that is needed for this action point to make
>any sense.)
>
>Regards,
>
>- Håvard
>
>  
>