From shane at ripe.net  Wed Oct 26 13:01:40 2005
From: shane at ripe.net (Shane Kerr)
Date: Wed, 26 Oct 2005 13:01:40 +0200
Subject: [whoisserver-dev] Re: Technical Question about Whois
In-Reply-To: <200510251428.j9PES82F003700@ns1.afrinic.net>
References: <200510251428.j9PES82F003700@ns1.afrinic.net>
Message-ID: <435F6214.4000705@ripe.net>

Frank,

[ I've asked Can to add you to the whoisserver-dev list that we've set
up to discuss details of Whois software development. I've also Cc'd the
list with this reply. ]

Frank wrote:

>I've got another question for you. How do the organisation tables work? I
>notice there are :
>
>| org                 |
>| org_name            |
>| organisation        |
>| organisation_id   
>
>How are these used ? How are the object_ids generated for them and how are
>they linked to the other tables?
>  
>
The org table is used for the "org:" attribute, which can appear in any
object.

The organisation and org_name tables are for the respective attributes,
which appear in the organisation objects.

The organisation_id table is used for the auto-generated "organisation:"
attributes. We base the attribute on the first letters of the
organisation name, plus a number. So "Shane's ISP" would have
"ORG-SI###-RIPE" as the "organisation:" attribute. The number, ###, is
determined by looking at the organisation_id table. So the code does
something like:

SELECT range_end+1 FROM organisation_id WHERE space="SI" AND source="-RIPE";

The source column is the same as the "source:" of the database: RIPE for
us, AFRINIC for you.

--
Shane