[419] in Hesiod
Re: info about hesiod
daemon@ATHENA.MIT.EDU (Greg Hudson)
Mon Sep 30 12:00:15 2002
From: Greg Hudson <ghudson@MIT.EDU>
To: Vivek Nallur <nvivek@ncst.ernet.in>
Cc: hesiod@mit.edu
In-Reply-To: <Pine.WNT.4.44.0209301505090.668-100000@Vajra.ncst.ernet.in>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: 30 Sep 2002 11:59:32 -0400
Message-Id: <1033401572.14743.96.camel@error-messages.mit.edu>
Mime-Version: 1.0
If you're specifically asking about the use of class HS for Hesiod
queries (as opposed to class IN), it's a mistake which has been
corrected at MIT and in the software distributed by MIT. We now use
class IN for Hesiod queries.
I've seen the use of HS attributed to a desire for administrative
separation; this justification is misguided, since DNS already provides
administrative separation through zone delegation. On the down side,
the use of a separate class requires special configuration on the
caching name server, causes incompatibilities with a lot of name server
software, and of course means that you have to replicate the class IN
delegation structure (locally or globally) for sites which use Hesiod.
If you're asking about Hesiod in general, I'm not sure what a good
source is. The ancient hesinfo(1) man page (should be in the software
distribution at athena-dist.mit.edu:pub/ATHENA/hesiod) describes the
tables we use it for, I believe.
I did find ways to download the 1988 Dyer paper pretty easily by
googling for "Hesiod Dyer". For instance:
http://citeseer.nj.nec.com/dyer88hesiod.html
Scanning it briefly, it looks like it's mostly accurate, except that it
talks about class HS and describes the old API. (The new API is
documented in man pages in the software distribution.)
