Standards Committee Teleconference 2008 08 13

From Project VRM
Revision as of 01:38, 14 August 2008 by Iain.Henderson (talk | contribs) (→‎Volunteered Personal Information)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Standards Committee Teleconference Call Notes

Drafted by Joe Andrieu, August 13, 2008

Other Calls

Category:Standards Committee Teleconferences

Attendees

  • Joe Andrieu
  • Iain Henderson

Previous Action Items

  1. Additional thoughts to one night stand notes to VRM wiki. Eve. By next Wednesday.
  2. Close loop with and Doc and finalize Oct mtg date. Joe. By next Wednesday.

Agenda

  1. Set Agenda
  2. Review previous Action Items
  3. Review Action Items

Notes

Volunteered Personal Information

This is a meta-layer for all sorts of data types. A general framework. Allows the individual to have their own Privacy Policy, which companies or governments would have to accept IF they wish to access this data.

Within that, you'll have multiple different types of data, each with potentially different policies.

5 or 6 standard policies, ala Creative Commons, human and machine readable.

At least 90 types of data, categories with potentially different types of privacy policy.

Iain sees about a year of a public process to generate a viable specification.

Recent develpments with Markey in US Congress re: Behavioral Data.

Behavioral data is definitely one of the 90 data types... looking at the Trust Index. Currently, the average trust index score today is about 55, but we expect users would require 70 or better. So, companies would need to behave significantly differently in their data handling and use to be allowed to access the VPI.

Trust index is more about compliance--users will still likely have specific requirements.

At the moment the current privacy legislation doesn't go to enough detail.

In a VPI share scenario, the iteration will be more akin to:

"Ok, you wish to access my XXX data, between this date and that date, and you are going to use it for four things and NOT these 23 things."

And there would be both technical parts to enforce these uses and policy parts.

Technical?

Either Higgins/Infocard or Liberty, in terms of access privileges.

A major issue is that once you release the data, you've lost control (with the old paradigm). The Liberty/Infocard/ Higgins technologies can enable seamless use of data with robust access privileges, bound with good policies.

Not traditional media "DRM" in terms of wrapping bits it crypto packets... but looking at a higher level. That said, DRM might be appropriate for very specific data types.

Right now, Iain is buying a new car. There are seven different car companies/dealers that have Iain in their sites, each with a % guess that he is in the market. So, Iain is stuck dealing with the deluge of junk mail, because there isn't a single repository for an authoritative source of his current intention that can release that under his terms.

The technology is really about access, policy, and auditing (APA). NOT about protecting the bits, but about building a framework for compliant access to the bits.

And discovery, machine readability. (Is this inseparable from the APA?)

So, there are two different components: one is about APA, the other is about interoperability so that the different data types can be seamlessly sent from user to and compliant vendor (and vice-versa).

Ultimately, there will be twelve components of the program. Of the twelve questions in the trust index, eight of them are from EU privacy law (which is pretty much the same as Canadian and Australian), and four are from VRM conversation.

How do we avoid the inevitable rabbit holes of debating data formats while establishing the framework?

So we don't need to solve all the data types at first. The CIC will be piloting this stuff from around January 2009. It doesn't need standards. It knows what data to offer up.

We want to make it clear that the APA needs a data-agnostic approach. Ideally, this would be something like MIME attachments for email, where MIME specifies how you attach any kind of data to an email, and the MIME types deal with individual data types and how they are encoded. Similarly, the APA framework deals with how one accesses any kind of data, and APA Datatypes would specify the formats and semantics of particular data.

So, in essence APA is like a VRM Mime Type for volunteer personal information, and Iain already has identified at least 90 APA datatypes.

[NOTE: terminology is all strawman. It made sense during the call, but may not fit long term.]

CIC will be looking for sponsorship

Soon, the CIC will be announcing a project to research VPI standard and will be talking with large companies to fund it.

This is separate from the Liberty specification. Liberty is the organizational vehicle for facilitating the creation of this specification.

Working with Liberty/NewOrg will allow access to resources, including personnel and wikis, etc.

SIG Charter draft has been sent to committee members.

Joe says it looks good. "I'll be joining Liberty to support this." I'm a bit concerned about IPR, but we'll work with that.


User-driven Search

Search maps are a form of VPI. They would be one of the datatypes.

  1. File format
  2. Interactive API (macros)
  3. Policy Standards (Data Rights Management)


Future Agenda Topics

  1. Discuss the distinctions between a Service Endpoint and a Service Manager.

Action Items

Next Meeting

Standards Committee Teleconference 2008 08 13