emv@ox.com (Ed Vielmetti) (05/12/91)
networks with multiple subdomains have a problem of keeping track of all those little domain abbreviations -- let's see, is the math department "math" or "mth" or "mps" or "mcs", what does the computing center call itself these days, etc. Trying to track down what is where is un-necessarily frustrating. this is a hard problem, since different organizations break down along different lines, and it's not always possible to tell in which of 2 or 3 or 5 different departments your colleagues might be in another organization. as a small aid to the problem, i point you to the uiuc.edu domain's use of "HINFO" records to spell out the full names of departments. e.g. cso.uiuc.edu. 50375 HINFO Computing_Services_Office NONE ag.uiuc.edu. 50400 HINFO College_of_Agriculture NONE To the extend that these tokens match with existing or eventual X.500 identifiers, so much the better. It's often possible to discover (or extract from a human postmaster) similar information and lists, but to the extent that more information can be sensibly slipped into the domain name service it seems a good idea to do so. -- Edward Vielmetti, vice president for research, MSEN, Inc. emv@msen.com "With all of the attention and publicity focused on gigabit networks, not much notice has been given to small and largely unfunded research efforts which are studying innovative approaches for dealing with technical issues within the constraints of economic science." RFC 1216
davy@ERG.SRI.COM ("David A. Curry") (05/13/91)
From: swrinde!sdd.hp.com!caen!sol.ctr.columbia.edu!emory!ox.com!ox.com!emv@ucsd.edu (Ed Vielmetti)
Date: 12 May 91 00:29:39 GMT
Subject: innovative use of HINFO records
as a small aid to the problem, i point you to the uiuc.edu domain's
use of "HINFO" records to spell out the full names of departments.
e.g.
cso.uiuc.edu. 50375 HINFO Computing_Services_Office NONE
ag.uiuc.edu. 50400 HINFO College_of_Agriculture NONE
To the extend that these tokens match with existing or eventual X.500
identifiers, so much the better. It's often possible to discover (or
extract from a human postmaster) similar information and lists, but to
the extent that more information can be sensibly slipped into the
domain name service it seems a good idea to do so.
An interesting idea, but isn't this what TXT RR's are for? (I realize
BIND didn't support TXT RR's until recently.) I've done something
somewhat different here:
I have put a TXT RR in for each and every host which says things about
the configuration of the machine ("diskless", "server", "dataless",
"standalone", etc.). Then I have a program which dumps the name
server, pings all the hosts, and then generates files for rdist based
on the contents of the TXT RR's and whether or not the host is "up" or
not. Works fairly well.
My idea for the TXT RR was to impose some sort of format on the string
(or use multiple records, I suppose) so you could go so far as to do
something like:
IN TXT "owner=comp. sci.;user=joe blow;room=EJ305;config=dataless"
or
IN TXT "owner=Department of Computer Science"
IN TXT "user=Joe Blow"
IN TXT "room=EJ305"
IN TXT "config=dataless"
or whatever. The only problem, at least in our environment, is that
every time a machine gets shuffled (which happens often around here),
the TXT RRs have to be updated if you keep location and user
information in them. This can get to be a pain real fast.
--Dave Curry, SRI Internationalemv@OX.COM (Edward Vielmetti) (05/15/91)
>> as a small aid to the problem, i point you to the uiuc.edu domain's >> use of "HINFO" records to spell out the full names of departments. >> e.g. >> >> cso.uiuc.edu. 50375 HINFO Computing_Services_Office NONE >> ag.uiuc.edu. 50400 HINFO College_of_Agriculture NONE >> >> An interesting idea, but isn't this what TXT RR's are for? (I realize >> BIND didn't support TXT RR's until recently.) I've done something >> somewhat different here: looks like a useful set of data, though as you suggest it can be a pain to keep up to date if system are moving around. still doesn't help answer the question of "what exactly is an erg in erg.sri.com", though. (engineering research group? one dyne-centimeter?) even trolling through the x.500 directory doesn't yield that information. -- Edward Vielmetti, MSEN Inc. moderator, comp.archives emv@msen.com "often those with the power to appoint will be on one side of a controversial issue and find it convenient to use their opponent's momentary stridency as a pretext to squelch them"
paul@uxc.cso.uiuc.edu (Paul Pomes - UofIllinois CSO) (05/15/91)
David Curry observes that TXT records may be more suitable than HINFO records for describing domains. At the time I started entering the records the TXT record was not supported. Even today it's not a RR that many people think to check when perusing another domain. For hosts I keep the HINFO information fairly generic. The CPU type is specified clearly (sun-3/50, vax-3500, etc) but the OS is kept generic (UNIX, DOS, VMS, etc). Version information just makes the information out of date sooner and gives a false sense of precision. /pbp -- Paul Pomes, Computing Services Office University of Illinois - Urbana
davy@ERG.SRI.COM ("David A. Curry") (05/17/91)
From: paul@uxc.cso.uiuc.edu (Paul Pomes - UofIllinois CSO)
Date: 15 May 91 14:32:55 GMT
Subject: Re: innovative use of HINFO records
David Curry observes that TXT records may be more suitable than HINFO
records for describing domains. At the time I started entering the
records the TXT record was not supported. Even today it's not a RR
that many people think to check when perusing another domain.
A valid point. Personally I usually use the "any" option when perusing a
domain, just to pick up such things (e.g. "dig erg.sri.com any"), but I
could see where lots of people don't do that....
I guess we just need to plead to the implementor-types for the "responsible
person" RR or whatever it's going to be called, and another RR for "owning
entity" or some such. Then we can stop all this stashing of info in random
places once and for all.
--Davepiet@cwi.nl (Piet Beertema) (05/17/91)
David Curry observes that TXT records may be more suitable than HINFO records for describing domains. At the time I started entering the records the TXT record was not supported. Even today it's not a RR that many people think to check when perusing another domain. True. But it's not so very long ago that MX records were a curiosity. They're commonplace nowadays. And I expect the same to happen with TXT records; they may well prove to be a valuable tool in finding real descriptive information about domains. For hosts I keep the HINFO information fairly generic. The CPU type is specified clearly (sun-3/50, vax-3500, etc) but the OS is kept generic (UNIX, DOS, VMS, etc). Version information just makes the information out of date sooner and gives a false sense of precision. Right. And if the information is up to date you may attract certain types who exploit the weaknesses and flaws of certain OS's. Sure enough, security through obscurity doesn't work in the end, but providing too much information is not a good idea either. -- Piet Beertema, CWI, Amsterdam (piet@cwi.nl)