[db-wg] Country Codes and that lack thereof

I have a number of very useful Perl scripts now that can and do parse
and then do interesting and useful things with RIPE WHOIS records,
and also with WHOIS records from other regions.

In particular, I have a tool that, if given a sizable IPv4 address
range, can show the names *and* ISO-3166 two-letter country codes
for all of the ASNs that are routing parts of the given IP range,
in a nice pretty table.

Mostly, this just works, but as I have learned, there are sizable
number of cases within the RIPE WHOIS data base where the relevant
home country for the ASN is not present in the WHOIS record for the
ASN, and where figuring out the home country requires a lot of
educated guesswork (e.g. if DO NOT see a country:  line but you DO
see the name "Sophia" in an address: field, then you can guess that
the home coutry is Bulgaria, and that thus, the apprtopriate ISO-3166
code is "BG".

I would like to know if I am the only person on the planet who has a
concern about this issue/problem.  If so, then I will just slink back
into my hole and be quiet.  If not however, and if others also feel
that every organization and resource record present within the RIPE
dat abase really should have a contry: line that clarifies where the
bloody thing should, nominally, be considered to exist, then I have
a good deal more to say on this topic.

First, I'd like to say ARIN seems to have solved this problem long ago,
in the case of their data WHOIS data base.  My Perl script has yet to
encounter a single ARIN WHOIS record for a number resource that does
not contain an explicit Country: line.  But the absence of country:
lines in RIPE WHOIS records is, I'm sorry to say, quite extensive.

Second, I'd like to ask the obvious question:  How did things get like
this in the RIPE data base in the first plac?

When a party is creating a RIPE WHOIS record for some resource, are they
allowed to enter essentially any bloody thing they like, including even
leaving out such important information as the home country where the
resource is based or where it will primarily be used?  If so, why?  Why
isn't there a form on the RIPE web site, which resource holders must fill
out in order to create a new record, and that simply does not allow the
user to proceed unless and until the country information has been provided?

This is not a debatable policy matter.  This is just common sense, I think.

And it isn't as if such a web form would be at all difficult to implement.

Thirdly, I'd like to just state for the record that in essentially all of
the cases I have sewen so far where RIPE WHOIS records lack country:
information, it is, in general, reasonably easy to infer such information
from other hints in the record itself, as noted above.  Thus, NCC staff
could... and I believe should... go through all of the records that are
currently missing this country: infromation and manually insert it
themselves, on behalf of the resource holder.  (This would be neither
hard nor wexcessively time consuming, as the number of records that
lack the country: information appear to me to be a tiny minority of
the total number of records within the data base.)

Lastly, because I have already found a lot of these problematic WHOIS
records that lack the country: information, and because I have already
taken the time to scrutinize then all, and make the necessary inferences
to determine the likely country codes, I would like to offer myself as
a volunteer assistant to the NCC staff for this small project, i.e. of
cleaning up the existing records and insuring that they all do have at
least some indications of the relevant country in each case.  I will be
only too happy to lend my assistance towards this worthwhile goal.

Of course, some troublemaker will undoubtedly now pop up here to claim
that reasonable inforences cannot be made about the primary country that
should be associated with any given resource, and/or that many resources
are in fact used in a manner that crosses national boundaries.

I disagree with the former point, based on my experience so far, and would
point out that all resource records are *supposed* to have *some* address
for the resource holder in them anyway, and whatever that mailing address
is, for each resource record, that mailing address can and should suffice
as the best approximation of the "home country" for the resource.

Likewise, even though a given resource may be used in a multi-national way,
I would argue that it would still be better to have *some* country indicated
for it, rather than none, as is currently the case in many records.  The
perfect is the enemy of the good, and none of this has to be perfect in
any sense.  A best guess for the county in each case where the resource
holder has not supplied that information is necessarily going to be better
than nothing in most cases.

I look forward to comments on this simple suggestion.  Quite simply, I am
suggesting that every resource record and every ORG record should include
a country: line which itself should specify an ISO-3166 standardized country
code, no exceptions.  Most already do, or else have a fully spelled-out
country names, and it appears to an outsider, such as myself, that the fact
that something slightly less than 100% of these records have this country
information is just a matter of sloppy DB maintenance which can be and
should be corrected.  I hope you will all agree.


Regards,
rfg