Re: [weirds] New Version Notification for draft-sheng-weirds-icann-rws-dnrd-01.txt

[Eric Brunner-Williams <[e-mail]>]

On 3/21/12 11:18 AM, Francisco Obispo wrote:
> Not really/necessary,

really? see below.

> Port 43 format is free text, and with >240 ccTLDs, gTLDs, RIRs, and others, it is easy to see how different those implementations would be...

i hope most everyone subscribed to this mailing list is remotely aware
of ...

(a) that 954 et seq is delightfully specified as
c: [::printable::]\r\n
s: .*\r\n <socket close>,

and

(b) there is more than one operator of a :43 server.

however, i'm not aware of the indistinguishability of necessity or
utility for human, or subhuman, parsing of data available at every :43
server.

if you're suggesting that the necessity and utility of a means to
parse the :43 return from the .tf (terres australes et antarctiques
françaises) server is indistinguishable from the necessity and utility
of a means to parse the :43 return from the .com server, i find it
hard to agree.

to generalize slightly, there exists correlations, both of trademark
necessity and utility, and law enforcement necessity and utility, and
some whois data sources and formats, and there also exist correlations
in some whois data sources and formats and registry backend operators.

so i suggest that the 80% rule is more likely to be useful than some
100% rule, and 80% of the use and variation amounts to a much smaller
number of servers than ~400.

> Perhaps ICANN has some details on a whois study.

at least one person contributing to this charter discussion has spent
years on icann whois working groups and is aware of their various
studies. the data may skew towards some correlations i mentioned
earlier, and overlook one or more of the ~400 sources of data.

-e