[Imap-use] UW's DNS-based approach vs. perdition/proxies

[Mark Crispin]

On Wed, 23 Apr 2008, Jeroen van Aart wrote:

> I understand, and it is obsoleted by http://tools.ietf.org/html/rfc1912, 

> which appears to also suggest a low TTL to not be a good idea.


In the general case, yes.  This is a not a general case.


> Even if no 

> "authorative" rfc suggests a TTL of less than say 1 day is a bad idea, I 

> would think that current bad practices of certain name servers can make it 

> problematic to depend on low TTLs.


I have not heard of any name server which fails to implement TTLs 
correctly.

Can you identify such a name server, or are you just speculating that 
there is such a name server?

If there is such a name server, then it is broken.  The way to deal with 
broken software is to fix it.  If the site and/or vendor refuses to fix 
it, there are ways to compel obedience.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.