[nc-whois] statistical 300 - notes.

[Thomas Roessler <roessler@does-not-exist.org>]

(This message is best read with a fixed-width font.  It's the basis 
for what I reported during the phone call.)

Here are some notes I took while looking at the set of 300 selected 
responses.  What I did was to try to isolate responses to individual 
questions, and read these in batch - possibly together with things 
such as category or main use of the whois.

(In order to do this, I just saved the file as plain text, and 
attacked it with standard Unix tools; grep & friends.  Probably adds 
some error margin to numbers when the file format fails, but the 
general picture should be correct.)



Remarks (q. 20+)
----------------

 - Privacy concerns for personal domains
 - information about commercial activity should be available
 - identify persons responsible of technical harm & illegal 
   activity; trace origin of SPAM; identify business
 - No marketing use/marketing use only with consent.  Multiple
   accounts of actual marketing abuse are given.
 - replace e-mail with central contact form in order to protect 
   against spammers.
 - one law firm: deomstrate bad faith (#1192)
 - #2070: Prove history of infringement, business pattern.  This one 
   claims that whois is in line with privacy laws.
 - multiple cases: track down where web site users come from
 - multiple responses: law enforcement should obtain court order in
   order to access data
 - "privacy means you have something to hide"
 - #1612: Problem solving should take priority over cutting down spam.
 - #2013: Suggests _national_ databases (by registrant's 
   nationality), following national data protection rules. 
   (Individual user.)
 - introduce an abuse contact
 - #2606 advocates pseudonymuous mail
 - one respondent considers whois to be an important link between 
   cyberspace & real world.

Accuracy
--------
  		<5%	5-25%	25-50%	>50%
 general public	60%	28%	7%	5% 
 IP		50%	39%	7%	4%
 general - IP	63%	20%	4%	5%

Explanation: "general public" is overall numbers.  "IP" is those who 
list IP among the most important use they have for whois. "general - 
IP" are those who do NOT list IP as most important use.

I was interested in the accuracy perception by intellectual property 
folks and others.


Some more numbers
-----------------

 Ever harmed?	151 no	135 yes

 Usefulness:	204 adequate
		 27 inadequate

 Uniform format and services?
 	        256 yes
		 25 no

 Searches on others data elements?
 		107 no
		178 yes

(TODO: What do they want to search for?)

 Commercial e-mail from service providers?
 		168 yes
		 95 no


 Sell contact info?
 		145 no
		113 opt-in
		 24 opt-out
		  4 yes

 Maintain bulk access provisions?
 		 85 no
		172 yes

 Extend bulk access (provisions??) to other TLDs?
		 92 no
		162 yes

 Change bulk access provisions?
 	        127 no
		121 yes

How to change them?  Most mention privacy, opt-in.  More rare: 
Demand for general whois service, "thick whois".  See also "sell 
contact info".

 Enhanced search capabilitieS?
 		151 no
		120 yes

(TODO: Cost?)


How to improve
--------------

 - accuracy, validity !!!
 - standard form, timely update
 - centralize! standardize!
 - fire NSI
 - ccTLDs: Some not informative enough, some too much personal detail.
 - punish bad data - domain name on hold, bad faith, cancel 
   registration.
 - Have ISPs update their customers' information to avoid stale 
   contact info on handles/role-accounts
 - regularly e-mail Tech-C ("ping")
 - don't use registrars as contacts
 - "life was so easy 4 years ago"
 - protect personal information
 - individual or company?
 - accurate information + accurate requester information:  Restrict 
   access with digital certificates, and record who requests data, 
   so abuse can be tracked.
 - offline records for law enforcement.  online records should 
   contain name only.
 - hold registrars responsible for accuracy
 - #2215: Enforce accreditation agreement.

Description of harm
-------------------

 - inaccuracy, misdirected mesages, can't identify perpetrator
 - "This site is horrible"
 - #2241: Wrongly listed as site's Tech-C; had to spend money on 
   legal defense.  Is this plausible? Can it be avoided in the future?
 - unpaid bills
 - outdated information


-- 
Thomas Roessler                        <roessler@does-not-exist.org>