If I might offer a comment on this from the data protection perspective...the concept of "sensitive" data is difficult.  It appears in the 95 EU Directive, applying to health data, religious data etc....but some data that is in those categories is not sensitive (eg the fact that I got a flu shot, like 49% of Canadians).  Other data that is not considered sensitive or is protected (eg the fact that a woman purchases a maternity vitamin at a pharmacy) could be harvested in our world of big data, and cause her employer to cease her employment.  So I dislike the term, although the concept of risk in disclosure of data elements is a good one.  The next problem one has to face is that end users usually have a very poor perception of the potential risk of data disclosure, so our policies have to reflect that risk from the perspective of the experts here assembled.  I understand only a couple of aspects of that risk....data and security and some regulatory issues.  I have a really hard time figuring out the data map, how far this stuff is travelling, who is harvesting it, what the secondary value-added service market looks like, who purchases those products.....etc.  I depend on you folks in the business to tell us that, or point me to some tutorial that will help be do the research.
cheers Stephanie

On 2016-12-09 10:50, James Galvin wrote:


On 8 Dec 2016, at 17:28, DANIEL NANGHAKA wrote:

I look at sensitive data as data that exposes a threat to the registrant -
this is also contact information.

This is probably too broad a definition.  I suspect a case could be made that most if not all data is a potential threat to a registrant because in our world of big data aggregation and correlation are always a threat.
This is a serious problem with data protection law these days, regulatory overreach.  While it is true that big data is a significant threat, nobody has come up with a satisfactory solution that I have seen.

I don’t have a suggestion just yet for how to restrict this but it is otherwise a possible starting point for discussions.

Jim





Other records like the NS records may not be sensitive data.

Third parties pick this data and use it for spam. This is the point where I
think restricted Access is important. Where there is need for the data the
registrant should be asked for consent that the data be shared.

Daniel


Regards
Nanghaka Daniel K.
Executive Director - ILICIT Africa / Council Member - FOSSFA / Community
Lead - ISOC Uganda Chapter
Mobile +256 772 898298 (Uganda)
Skype: daniel.nanghaka

----------------------------------------- *"Working for Africa" *
-----------------------------------------



On Thu, Dec 8, 2016 at 10:11 PM, James Galvin <jgalvin@afilias.info> wrote:



On 7 Dec 2016, at 9:55, Greg Aaron wrote:

In the coming discussions, one approach could be: There are good reasons
to publish the thin data … is there any compelling reason _not_ to publish
it?   If we can take care of this low-hanging fruit, we will solve part of
the puzzle and we can concentrate on the issues around contact data.  This
is not a proposal to publish thin data only.  It’s an attempt to
disentangle concepts and find a way forward.  Not all data is the same, so
let’s stop treating all data the same.  We may not have to iterate
repeatedly about thin data.


I agree with the principle that we should tease apart “registration data”
into a few different categories.  The discussion in the rest of this thread
has been focused on that and I’ll state I support it.  My current view is
that there are at least three categories of data: PII (e.g., contact
information), operational and explicitly not-PII (e.g., registrar ID and NS
records), and other (e.g., registries with specific requirements).

I have two concerns with this discussion though.  First, we keep talking
about “publishing” data.  Greg is careful to point out he’s not talking
about publishing, per se, but he doesn’t mention what we are talking about.

Second, given we understand our data (which is a reason to categorize it)
there are at least three topics to talk about with respect to that data.

1. Why do we care about this data, or perhaps, what is the purpose of the
data?  The answer to these questions is both critical and essential.  They
will drive the answer to the next two questions.  In my opinion, without an
answer to these questions (eventually, if not first or early in our
process), discussions about the next two topics will never come to a
conclusion.  By the way, also my opinion, any answer that somehow embodies
a reference to the existing system and service is irrelevant.

2. What are the data collection requirements?  This includes who, what,
where, why, and how, including storage.

3. What are the publication requirements?  Might be zero.  Greg suggests
above that we could approach the problem of publication in some cases by
answering the following question, “Is there any compelling reason not to
publish it?”  I will object to this.  This is never the right question.
The right question is always, “Why publish it?”  You can’t publish it if
you don’t collect it and you don’t collect it if you don’t have a need for
it.

All questions must be answered in the positive.  Otherwise what’s the
point of our discussion?  The answer will always default to be collect
everything and publish everything, and then let the lawyers fight about
what’s public.

Jim

_______________________________________________
gnso-rds-pdp-wg mailing list
gnso-rds-pdp-wg@icann.org
https://mm.icann.org/mailman/listinfo/gnso-rds-pdp-wg



_______________________________________________
gnso-rds-pdp-wg mailing list
gnso-rds-pdp-wg@icann.org
https://mm.icann.org/mailman/listinfo/gnso-rds-pdp-wg