Re: Comments on <draft-cooper-privacy-policy-01.txt>
Thanks for your comments. Responses inline.
On Jul 8, 2010, at 11:05 PM, Bob Hinden wrote:
No hats on, these are my personal views.
I have now read the draft. My overall comment is that I am not
convinced if this is needed and am sympathetic to the views
expressed on the mailing list that this is solving a problem the
IETF doesn't have.
much simpler and of the form where it first starts with a general
statement that the IETF does it's work in public and almost all
information information supplied to the IETF is made public and will
be available on the IETF (and other related) web sites.
A simpler intro with a focus on the public-ness of the IETF is
I would then list the exceptions. For example, credit card
information for meeting registration and social tickets, and
information for "letters of invitation". Note: As I read the
draft, there is very little that actually falls into the private
category. This leads to to wonder about the scope of the problem
this draft is solving.
I tend to think that privacy risk isn't so much about the percentage
of sensitive data collected as about the sensitivity of any data
collected. The IETF interacts with credit card numbers, passport
numbers, authentication credentials, and other kinds of data that are
widely perceived to be sensitive. I think those deserve documentation
(as do less sensitive data elements like web logs, but I can
understand why others may disagree on that point).
The IETF goes to great length to tell people about how we do our
work and what is considered a public contribution, via the Note
Well. I would be surprised if anyone thought otherwise. Doing our
work in public is essential to how the IETF works.
I have issues with the Introduction. The first sentence says:
In keeping with the goals and objectives of this standards body, the
IETF is committed to the highest degree of respect for the privacy
IETF participants and site visitors.
This makes it sound like the highest priority of the IETF is
Privacy. I don't think this is true as I described above. The vast
majority of what the IETF does in Public. There is very little that
is Private. The IETF is careful about what needs to be kept private
and does not disclose it.
That sentence was cribbed straight from ISOC's policy and could easily
The Introduction says:
This policy explains how the IETF applies the Fair
Information Practices -- a widely accepted set of privacy principles
 -- to the data we obtain.
I don't know if it is appropriate that the IETF apply these
practices. Or if there are other practices that would be more
I know that the IETF is different from many other organizations, but
the Fair Information Practices form the basis of more or less every
information privacy law, regime, policy, best practices, self-
regulatory framework, and guidance document around the world. The IETF
doesn't have to reference them, but I think the reference makes the
document better rather than worse -- at least we're basing it on some
The IETF is different from other organizations in that much of our
data is public and not private.
It might make sense to remove the parts of the document that discuss
public data so that it only focuses on private data.
The rest of the Introduction appears to be a summary of the first
One suggestion I got on the -00 was to summarize the Fair Information
Practices up front since they may not be familiar to many people. So
the summary was by design.
 Organization for Economic Cooperation and Development, "OECD
Guidelines on the Protection of Privacy and Transborder Flows
Personal Data", http://www.oecd.org/document/18/
I don't know anything about this web page, who produced it, how
stable it is, etc, etc.
On who produced it, I think it's fairly obviously produced by the OECD.
It is fairly long, around 21 pages. I don't know if this is
appropriate for the IETF. I think it would better to not include
this information as it is hard to judge how appropriate it is.
Also, some of the practices seem to be at odds with normal IETF
practices. For example, it implies that individuals have complete
control of the data the IETF makes public. This isn't true in most
Removing the public data parts of the policy might help here.
Section 2 and 3
A lot these section is a summary of what is defined in other places
(References 2, 3, 4, 5, 7, 8). Other parts of the text are fairly
generic, such as the information that a web server can learn about a
web client. Not thing very IETF specific here. I don't see very
much value repeating this.
One benefit would be having all of it in one place, especially if this
turns into a layered policy that is referenced from a central location
The first paragraph:
The IETF does not sell, rent, or exchange any information that we
collect about our participants or site visitors. However, we will
disclose information under the following circumstances:
The first two "sell & rent" is true, but the "exchange" is not true
as you state later in the section. Much of the data we collect is
Correct. This was more language pulled from ISOC's policy that I can
I am not really qualified to comment on the specifics here, such as
how long credit card or letter of invitation information needs to be
retained. I would have thought that all financial data needs to be
kept for some number of years.
This describes our current operational practices regarding log
files. Including specific times for retention will make it hard to
change this in the future.
If the written policy is not too difficult to change, the actual
policy shouldn't be hard to change either.
Also, if log files are going to be covered, what happens to the
backups? Are we required to scrub the backups? This would be
difficult and expensive. What about backups of credit card
I need to find out the back-up policies, but the idea right now is
just to document the current policy, not change it.
In the acknowledgment section you cite the IAOC. The IAOC has not
done any formal review of this draft. It is better if you cite the
people in the IAOC you have discussed this with you and not list the
Now that I have written this, you can cite me if you choose :-)
I think most of the references are Normative, not Informative. That
is, this draft depends on these documents.
Fine by me.
Ietf mailing list