Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!umich!samsung!brutus.cs.uiuc.edu!psuvax1!psuvm!IVORY!MWH
From: mwh@IVORY.EDUCOM.EDU (Michael Hrybyk)
Newsgroups: bit.listserv.nodmgt-l
Subject: Re:  Routing tables from pathalias
Message-ID: <NODMGT-L%90020911214629@BITNIC.BITNET>
Date: 9 Feb 90 16:25:00 GMT
Sender: Node Management Discussion <NODMGT-L@BITNIC>
Reply-To: Node Management Discussion <NODMGT-L@BITNIC>
Lines: 25
Approved: NETNEWS@PSUVM Gateway
Full-Name: Michael Hrybyk
Comments: To: NODMGT-L%BITNIC@pucc.PRINCETON.EDU

Mike G. and I agree wholeheartedly here. My additions to pathalias
include a generic postprocessor, obviating the need for new software
releases every time table format changes occur in the network.
Format changes, site tailoring, and other issues *that have nothing
to do with general network routing* are thereby kept separate, and
are relegated to a policy decision (what should the config file
look like, what information to include in a routing table, ...).

As for the amount of work involved:

1. the pathalias core was already done.
2. the NODES parser front end was needed for other tools (hasn't everybody
   done this? why don't we standardize some access routines in C, Pascal,
   REXX, ...?)
3. the postprocessor was an add-on, and would seem to cut down on
   the amount of maintenance needed later. This produces a net savings
   in time over the life of the code. It also eliminates the need
   for a set of programmers per OS developing system-dependent post
   processors.

Of course, one can still generate a generic table that some arbitrary
post-processor can transform.

Mike Hrybyk
BITNIC