:: Re: [DNG] Weird network issue - slo…
Página Inicial
Delete this message
Reply to this message
Autor: Rick Moen
Data:  
Para: dng
Assunto: Re: [DNG] Weird network issue - slow to resolve IPs [solved]
Quoting Steve Litt (slitt@???):

> I just got thru updating my unbound doc at
> http://troubleshooters.com/linux/unbound_nsd/unbound.htm , and added a
> new glossary. This is very necessary because there are so many
> contradictory terms for things over the Internet. I'm not saying my
> glossary is the most accurate resource on the Internet, but I think my
> glossary probably does the best job of de-contradicting terms people
> throw around. I spent three weeks writing this glossary. It was one of
> the toughest docs I ever wrote: Tougher than any one chapter in Samba
> Unleashed.


I sympathise with you needing to do all of that painstaking work, Steve.

In the glossary, you take particular pains to declare bogus for purposes
of your guide the phrase 'iterative server' and similar. But I cannot
recall seeing that phrase, so I'm unclear on why you needed to say that.

In my 'Village of LAN' piece, I speak in one of the examples about 'a
nameserver offering iterative service only', and in several other places
use the subphrase 'iterative service' -- clarified elsewhere in the
piece in contrast to recursive service. But it would not have occurred
to me to use the odd phrase 'iterative server'. It's not the _server_
that iterates, but rather the query, for lack of ability to send back an
answer properly with the RA = Recursion Available bit set when sent a
query with the RD = Recursion Desired bit set.

My point is, you seem to be walking the long way around a pothole that
wasn't actually there to fall into. IMO.

Although I don't share your desire to use the word 'resolver' in the way
your document does -- because in context it ends up creating too much
confusion -- you're of course correct that that usage is observed in DNS
discussion.

And that reflects, as you say, there being too many decades of sloppy
and gradually changing terminology. I don't _think_ that problem's as
bad as you suggest, but it's definitely present.