Re: another lookup problem

["David R. Conrad" <]

Hi Wilfried,

>=To what end?  Is the purpose of the database registration information
>=lookup or a more general white pages service?  I' would suggest the
>=former and leave the latter to the WHOIS++/LDAP/X.500 crowd.
>  
>  - local langauge problems (European National Character Sets) which make
>    it difficult to find out about a (arbitrarily chosen) transcription.

Local language problems?  Heh.  Let me introduce you to traditional
Chinese with all 100,000+ characters :-).  The advantage of the whois
database serving as a simple registered object lookup mechanism is
that you don't have to worry (too much) about gunk like local language
support -- the registration database access method (whois) allows
lookups on registry objects, namely inetnums, handles, domains, route
objects, etc. of which people (generally :-)) are not a part.

Yes, I'm arguing person: field contents should not be indexed, but
I've been known as a radical in the past.  Note this would have
solved the problem that brought this issue up (I think).

>  - comments from other environments, where the lack of this feature was
>    quoted as a major obstacle to use this software 'cause this
>    functionality is provided right now.

If all you have is a hammer, everything begins to look like a nail.
This doesn't mean you should attach a blender to your hammer to allow
making more consistent puree...

>  The remark about the white pages is understood, although I don't think
>  we should limit the sw functionality to avoid misuse :-)

I'm of the "simpler is better" camp.  It would be nice if registry
objects were defined to be strictly hierarchical in nature (addresses
and domains are already there, handles are soon to be there with the
addition of the -<registry> suffix (APNIC's will be -AP (not -APNIC)
as soon as I can find time to write the necessary software to do the
conversion), AS numbers are somewhat problematic, but there are only
64K of them and we can make RDIs hierarchical when IDRP gets out of
the starting gate).

This would allow a simple, programatic way of determining which
registry maintains which objects (I'd like to use the DNS to do this,
but rwhois is still lurking in the background).  I figure it is the
only way the registry database system can scale and keep up with the
Internet growth rate.  Flat namespaces, like people's names can be
handled by other methods (I'm told centroids work, but don't know
enough to have an opinion).

Cheers,
-drc