[nc-transfer] Some notes...

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

[I hope this makes it to the list.  If not, I'd like to ask Dan 
Steinberg to forward it.]

As promised, here are some notes on what I said during today's 
conference call.  Please note that all these are just suggestions 
 from a non-member of this task force.


I. Survey Design

 - Basic principle: Avoid free-form, use multiple-choice.  Make sure 
   you get a survey which doesn't cost you months to evaluate.
 - Keep wording simple and understandable.  Make sure that choices 
   are clear, and mutually exclusive (where necessary).
 - Free-form is fine in order to gather some nuggets, but not in 
   order to gather "hard facts".
 - Ross Rader suggested to give an e-mail address for comments 
   instead of providing comment space on the survey form.  I didn't 
   make that remark during the call, but I think the idea is 
   excellent:  It adds an additional burden to the respondent, so we 
   may hope that only those who really have something interesting to 
   tell will make use of that option.   I'd expect few, but quality 
   responses.
 - Some answers will have to be broken down by the registrar used.
   Implementation:  Ask for registrar(s) first, then produce form where 
   registrar names are already filled in, and present this form to 
   respondent.
   Problem: Many users may not know which registrar they are using.
 - Christine Russo suggested to also ask about specific policies 
   (auto-ack, auto-nack, asking for confirmation, ...).  Please note 
   that this is complementary to asking for experience.  Things 
   which may sound good in theory may work out badly in practice.
 - In connection with the question about which policies registrants 
   are talking, it was suggested (not by me) to ask for _recent_ 
   experiences.  I believe that this is a very good idea.

II. Registrars' statistics

 - I suggested to use registrars statistics in order to understand 
   what kinds of respondents we have.  In particular, statistics on 
   failed and successfull transfers, and on the reasons why transfers 
   were rejected, may help to detect anomalies among respondents. 
   This may help in order to detect large amounts of fraudulent
   responses; note, however, that I'm not entirely convinced myself 
   that this will be entirely successful.  It may still be worth a 
   try.

 - What would be even more interesting would be the development of 
   transfer statistics over time:  If my guess is right, the number 
   of transfers per day (or week) will be a rather smooth curve, 
   with bumps happening at the point of time of policy changes. 
   Comparing the effect of policy changes on such statistics as 
   failed transfers, complaints about fraud, and the like, may help 
   the task force when it tries to understand the impact some of the 
   policy changes actually have had.  Note that this kind of 
   approach would require quite a bit of data.  Just snapshots would 
   most likely not be enough.  

   Once again, this is just a suggestion.  Feel free not to implement
   it, in particular if you believe that the cost involved would not 
   be justified by the possible impact on policy development.


III. Some suggestions for specific questions.

Here are some suggestions for specific questions which could be 
asked.  Note that I have formatted them with just a single answer, 
for the sake of simplicity. In a final questionnaire, I'd imagine 
that these would be treated like the "number of attempts" example I 
gave in the edits to the draft which were posted to this list.

 - Have any requests for transfer you made been turned down?  If so, 
   what reasons were given?
   
   [] Dispute Resolution Policy  [what precisely is stated there?]
   [] Pending bankruptcy of the SLD Holder
   [] Dispute over the identity of the SLD Holder
   [] Request to transfer occured within the first 60 days after the 
      initial registration with the Registrar
   [] Domain was in unpaid status
   [] Other:
   
This is Exhibit B + unpaid, which are, according to my recollection, 
the most frequent reasons mentioned in the complaints posted to 
various mailing lists.
   
 - (Replacement for q. 11a; choices need additional work.) If the 
   losing registrar contacted you about a transfer request you made, 
   what information did their message contain?
   
   [] Advertising material on the losing registrar's services.
   [] Renewal instructions.
   [] Instructions on how to complete the transfer.
   [] An indication why they contacted you instead of completing the 
      transfer.
   [] Other:

(Anything else?)

 - If the message you received contained instructions on how to 
   complete the transfer, were these instructions comprehensible, 
   accessible, and complete?
   
   [] Yes
   [] No

 - If both renewal and transfer instructions were given, which 
   instructions were more comprehensible, accessible, and complete?
   
   [] Both were equally comprehensible, accessible, and complete.
   [] Transfer instructions were more comprehensible, ...
   [] Renewal instructions were more comprehensible, ...

 - How much time did you have to complete the transfer after receipt 
   of the message from the losing registrar?
   
   [] 1 ... 5 days

 - If the losing registrar contacted you about a transfer request you 
   made, and you fulfilled any steps you were asked to perform in 
   order to complete the transfer in a timely manner, did the 
   transfer succeed?

   [] Yes
   [] No

 - What reason, if any, did the losing registrar give for contacting 
   you about your transfer request?

   [options TBD]



-- 
Thomas Roessler                          http://log.does-not-exist.org/