[ppml] Directory Services - Take 2

[william(at)elan.net]

On Fri, 10 Jun 2005 bmanning at vacation.karoshi.com wrote:

> 	maintaining servers and updating data are not the same
> 	as a supported and maintained protocol... which was
> 	Michaels claim.

Which is exactly what I meant and you appear to have missed.
NEITHER whois NOR rwhois protocols are maintained!

But comparatively speaking for those unmaintained protocols the use of 
whois is bigger on the order of 1000:1. So not surprisingly there are 
calls for ARIN not to support rwhois (but no calls I hear of the same
for whois!).

But really we can't stop supporting either one until we have adequate
replacement and rwhois is more advanced protocol then whois, so whois
is NOT replacement for rwhois and calls for ARIN to stop supporting it
are wrong and we don't have anything better standardized or implemented
to enough of the degree that we could even say that using rwhois
IS NOT RECOMMENDED (take it as IETF words). Now once IRIS is out of the 
ietf standardization bag and begins to get some use than in maybe 2-3 
years we can re-examine that and consider saying that RWHOIS support
should be shifted out.

> 	tweeking data formats.  It is instructive to go read the actual
> 	WHOIS specs.  So no, I'm not confused.

The great thing about whois is you don't need to read the spec to implement
it (and that is coming from somebody who has definitely read the spec many
times just in case), its just TCP ask/answer with no serious effort to 
standardize the functionality or format.

So if we agree that whois is just all that and any attempts to introduce
format into the response is bad, then lets go ahead and quickly dump RPSL
since it apparently standardizes the wrong thing and impedes on somebody
else's protocol space! Of course its not all that easy because sometimes
protocols are built on top of the other, we have seen it many times -
remember earlier post by Michael about SOAP and XML-RPC? LDAP has also
been used in that way and in fact had been seriosly considered as basis 
for future whois at CRISP WG as I mentioned.

P.S. If you look at HTML it is in fact tweaking of the old format, they've
been rather successful at that too...

-- 
William Leibzon
Elan Networks
william at elan.net