[db-wg] whois_rip feature request, multiple inverse attribute queries needed

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Shane, all,

Shane Kerr wrote:
> Problem 2: PERSON searches

[...]

> The database has no way to know if you are looking for the "nic-hdl:" or
> the
> "person:" attribute that matches.
> 
> 
> Possible Solution A: multiple inverse queries
> 
> We can solve problem 1 (multiple "origin:" attributes), and Tobias'
> original
> problem, by allowing multiple inverse queries in a logical AND operation.
> We'd need to change the meaning of multiple "-i" flags though. 

[...]

> Limitations are:
> 
> o This does not work for problem 2 (ambiguous "person:" and "nic-hdl:"
> attributes).

This would not be a problem when looking for non-ambiguous attributes
like a handle. I'd like to guess that such a query is performed in most
cases by searching a handle and i.e. a mntner instead of a real name and
other attributes. Don't know if this is possible to implement, however...

> o It is not exactly simple for the average user to understand.

If the other, simple query type with just one -i option continues
unchanged, there would be no difference to anybody not interested in
advanced queries.

> Possible Solution B: SQL-like searches
> 
> While we would be reluctant to allow actual SQL searches into the database,
> we might be able to open the database up to an SQL-like search language:
> 
> $ whois --search tech-c='TOC76-RIPE' and mnt-by='CW-EUROPE-GSOC'
> 
> $ whois --search nic-hdl='kiri'
> 
> Problems:
> 
> o One could perform some extremely non-optimal searches. (We could possibly
> avoid most of this problem by requiring that some key is present in the
> search.)

It would be an affordable disadvantage being forced to use specific
attributes, im my opinion - depending on what attributes that will be,
of course :-)

> o A second query language for users to learn.

Same as above, isn't it? It's an advantage for those how need this
feature and doesn't hurt those that only need "simple" queries.

It's a second query language to learn anyway, both if implemented in the
RIPE Dbase software and if done in CRISP, if I get that right.


Regards:

Tobias



- --
Tobias Cremer M.A.
IP Admin Engineer

Cable & Wireless Telecommunication Services GmbH
Landsbergerstr. 155     80687 Muenchen  Germany
Tel +49 89 926 99 0 --  FAX +49 89 926 99 180  -- COMNET 7 49 9169
www.cw.com/de

- -- Every message GnuPG signed --
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFCFILJhC6y11CNwvcRAn+IAKDWZjBtvBqgGkKxXzSG5Uo81GXupACfZjY4
q4cABiJ+xs5xl58Dv0Qidq0=
=IXZe
-----END PGP SIGNATURE-----