--On Thursday, April 03, 2003 10:04:37 -0800 Tony Hain <[EMAIL PROTECTED]>
wrote:

> This whole deprecate effort is about making it someone else's problem.

It is less of a problem to do the right thing, as Elliot wrote. Less pain
now, less pain in the future. 

> The IETF has successfully driven out most of the network managers, and
> now the remaining developers are trying to dictate how networks get run
> by removing the tools that people really use. 

You are right in the sense that there are too few operations people in the
IETF standards work. And yes, the problem is that the remaining developers
are inventing things we as operators (I help run a number of smallish
research nets, only OC-192 lines and perhaps 2-300000 users, not much to
talk about, but I think I count as "operator".) do not want. 

Your analysis and argument fails on the pivotal point, though: 

We do not want ambigious addresses. The only reason to do that is were
there lack of address space. Really. 

We do not want NAT. Nat breaks things. Not for web surfing, but most other
things require hacking, be it Kerberos tickets without address info or what
have you. 

We do not want unreachability built-in. We can do that ourselves, or by
means of backhoe. 

We do not want to break DNS. It is in a enough sad state as it is. 

I am talking to a lot of friends in the operations community these days, to
make them aware of this consensus call, and have them say their meaning.
Without exception their assessment of SL can be summarised to "OUTRAGEOUSLY
UGLY". 

So, I'm glad that you call on the operator community, but I think you will
be surprised at what they say. 

-- 
M�ns Nilsson            Systems Specialist
+46 70 681 7204         KTHNOC  MN1334-RIPE

We're sysadmins. To us, data is a protocol-overhead.

Attachment: pgp00000.pgp
Description: PGP signature

Reply via email to