W3C home > Mailing lists > Public > public-privacy@w3.org > October to December 2012

Re: privacy definitions -- was: WebID questions

From: David Singer <singer@apple.com>
Date: Fri, 19 Oct 2012 08:11:23 +0800
Cc: public-privacy list <public-privacy@w3.org>, WebID XG <public-xg-webid@w3.org>
Message-id: <CCF9B9D5-D181-4D68-9799-D95A37674EE1@apple.com>
To: Kingsley Idehen <kidehen@openlinksw.com>

On Oct 18, 2012, at 19:54 , Kingsley Idehen <kidehen@openlinksw.com> wrote:

> 
> I am talking about "identity" and all of the constituent aspects relevant to browsers.

OK

> 
> Cookies are identifiers,

I think you are missing 'some':  'some cookies are identifiers'.  Cookies are simply a piece of state that the service records; they happen to use the user's disk to do that. Some cookies are used to record an online identifier, some are not.

> they are part of the "identity matrix" .

I'm sorry, a cookie that records 'this user has visited before', or 'this user has a packet of biscuits in their shopping cart' are only part of an identity matrix to the extent they contribute to a fingerprint.

> Privacy is ultimately about self calibration of one's vulnerability. This fundamental principle is realm agnostic.
> 
> A browser should put me in position to know:
> 
> 1. my current login identity at any given point in time

But 'login identity' is not known to the browser.  We could work on the protocols to expose it, but there is nothing formal in forms (at the moment, but see below) or cookies that clearly identifies which form fields or cookies constitute identity management.

> 2. nature of all identifiers used to construct my identity -- this includes cookies.

This *might* (often does) include *some* cookies.  Cookies are not the only way to communicate identity, and not all cookies are involved in identity.

> The browser has all the data,

No, as noted above, it does not.  There is no 'role' or 'purpose' associated with a cookie.  Browsers have no idea what the state is.

> it just needs to improve the UI/UX based on a better understanding and appreciation of identity, in a manner that's fully controlled by the end-user.

That 'just' hides a whole load of things.

> We can only start solving this problem by accepting the fact that end-users should possess full control over their identity when working with browsers.

You seem to be setting up a strawman here that the only reason this isn't happening is because people disagree with this principle.

> We should be rapidly moving away from today's surreptitious identity patterns which operate of the false assumption that the browser application developers know best, when in fact they don't.

I'm sorry, again, you are attributing positions to browser application developers which I don't think they hold.

> Identity is inherently nebulous. Thus, it's always best controlled by the Referent (ultimately an end user) of identifiers.

Phrases like 'always best' are quite sweeping; are you quite sure no-one can think of a counter-example?


You should know that I argued at the 'identity on the web' meeting that we *should* label form fields that create or quote user-ids and passwords ('this field is where the user supplies a suggested user-id', 'this field is where the user supplies their online password', 'this field is where the user answers a secret question') and so on, partly so that some of the things you ask for could be enabled - the user-agent can 'watch' your online account creation and login/logout behavior and manage more than just passwords for you.

I guess what I am saying is that you seem to be assuming an adversarial position when I don't think it helps or is needed.

David Singer
Multimedia and Software Standards, Apple Inc.
Received on Friday, 19 October 2012 00:11:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 19 October 2012 00:11:59 GMT