Re: [enum-wg] 9.3.e164.arpa down

[Jim Reid jim@localhost]

On Nov 16, 2006, at 15:34, Stastny Richard wrote:

> Jim, you are wrong here. The "Interim Procedures" state nowhere that
> they are "only for trials"
I didn't say they did Richard. SG2 agreed the interim procedures on  
the understanding that they were for trials and not for production  
service. Certain ITU members would not have allowed the procedures to  
go through unless that condition was attached. We were both at the  
SG2 meeting where those procedures were adopted. Perhaps we have  
different recollections of that meeting?
> OTOH, you approach to simple state that sh*t happens is also
> not feasable. IMO RIPE should first look what can be done and
> Michaels proposals should at least be taken into account
I didn't say that either Richard. And I didn't discount Michael's  
constructive suggestions at all. The opposite in fact. If software  
does not behave reasonably on encountering lame delegations -- a  
depressingly common DNS problem -- that software needs to be fixed.  
Not that that implies lame delegations should be tolerated. Or that  
the administrators of those delegations shouldn't be contacted when  
these problems are found.
This thread started with a complaint along the lines of "It hurts  
when SIP servers can't cope with lame name servers for 9.3.e164.arpa.  
Let's pull the delegation." [I paraphrase and exaggerate for  
effect.]  I pointed out why this was flawed reasoning, politically  
and technically.
I also encouraged Michael to do two things. One was to submit a draft/ 
paper to this WG on good DNS practices. If the WG picks up on that --  
ie they get something to actually work on -- there's then a  
"standard" for Tier-1 operators and bureaucrats to follow. Next, if  
there's something that needs to be done to make SIP servers more  
robust to lame delegation errors, these should be written up too.  
That piece of work may be out of scope for the RIPE ENUM WG. Though  
this is for the WG to decide.