W3C

Web Security Context Working Group Teleconference

30 Jan 2008

Agenda

See also: IRC log

Attendees

Present
Thomas Roessler, Maritza Johnson, William Eburn, Jan Vidar Krey, Tyler Close, Anil Saldhana, Bill Doyle, Ian Fette, Yngve Pettersen, Luis Barriga, Daniel Schutzer
Regrets
Johnathan Nightingale, Tim Hahn, Mary Ellen Zurko, Hal Lockhart, Serge Egelman
Chair
Thomas Roessler
Scribe
Ian Fette

Contents


<trackbot-ng> Date: 30 January 2008

<scribe> ScribeNick: ifette

Convene

tlr: f2f next week, see mez's agenda email
... some draft text low hanging fruit, take a look
... look at editor's draft, and follow-up via email re: discussion on what can be fast-tracked
... today is poor attendance day
... mez, tjh, hal, phb, et al not here
... current AGENDA
... quick look at wsc-usecases, then ISSUE-131, then ISSUE-130, then ISSUE-114
... and ISSUE-129

approve minutes

<tlr> http://www.w3.org/2008/01/23-wsc-minutes.html

RESOLUTION: minutes approved

wsc-usecases updates

tlr: no closed action items
... tyler, can you give us an update?

tyler: in tracker, only one action open against note to incorporate tjh's review
... that action is marked pending review, expecting tjh to go through doc and make sure it's good
... also got an email from Al Gillman, problem when reading note using screen reader. The table doesn't come out well
... he has people who will contact tyler re: how to reformat that
... want to turn table into list, tyler not so keen on because it's not so good pointing out what isn't covered

tlr: what table?

tyler: at the top of the list of scenarios is a table

tlr: takeaway is there is some editorial stuff left, tlr happy to provide help if he can

<tyler> http://www.w3.org/2006/WSC/drafts/note/Overview.html#scenarios

tlr: does tjh know that he's on the hook to review?

tyler: pinged mez, she put it into pending, expecting mez to ping tjh. tlr will email tjh

<tlr> ACTION-376

tyler: it's ACTION-367

<tyler> http://www.w3.org/2006/WSC/track/actions/367

tlr: all for this agendum

ISSUE-131: Executing code outside of browser

ISSUE-131?

<trackbot-ng> ISSUE-131 -- Executing code outside of browser in 8.3.2.3 is vague / scary -- OPEN

<trackbot-ng> http://www.w3.org/2006/WSC/track/issues/131

<tlr> http://www.w3.org/2006/WSC/drafts/rec/rewrite.html#techniques-robustness

tlr: some level of agreement in december, on some text

<tlr> http://lists.w3.org/Archives/Public/public-wsc-wg/2007Dec/0116.html

tlr: see this email
... lengthy discussion thread re: plugins and how all of this applies to plugins
... should we go forward taking effectively a mix of the text that Mez had proposed in Dec. with the change that Mike had proposed, or...?

ifette: i have no problem with the text in the email that the link tlr posted contains

tlr: reads text

yngve: would like to see the new version in written form
... should post somewhere

ifette: already there

tlr: action item to send mail to list

<scribe> ACTION: tlr to send email to list regarding ISSUE-131 containing full text of new proposal, and will close out the issue [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action01]

<trackbot-ng> Created ACTION-380 - Send email to list regarding ISSUE-131 containing full text of new proposal, and will close out the issue [on Thomas Roessler - due 2008-02-06].

tlr: assuming nobody objects to the email sent out, assume email ready to put to bed

ISSUE-130: Trust Anchor Consistency Across Devices

ISSUE-130?

<trackbot-ng> ISSUE-130 -- Trust Anchor Consistency Across Devices? -- OPEN

<trackbot-ng> http://www.w3.org/2006/WSC/track/issues/130

<tlr> http://lists.w3.org/Archives/Public/public-wsc-wg/2007Nov/0015.html

tlr: latest proposal from luis posted

<tlr> http://www.w3.org/2006/WSC/drafts/rec/rewrite.html#tls-across-devices

tlr: affected text is here (link above)
... content doesn't really control the UX
... we should not ask for the same UX
... suggests web content should offer trust and TLS consistency across UAs
... why doesn't luis introduce last changes

<tlr> http://lists.w3.org/Archives/Public/public-wsc-wg/2007Nov/0015.html

luis: only minor changes to proposed text, just clarifications
... first one is about user experience
... according to original text, says content should be designed to offer same UX, that's difficult, propose to change this, only thing web designers control is trust and TLS consistency, UX is on client side

tlr: so let me understand
... by the same argument, tls consistency is a UA question as well
... after all, the UA selects the key of trust roots
... direction is to say that having the same security experience is a goal which designers should aim for
... disinclined to say TLS and trust anchor consistency
... suggests for first part, instead of "designed to offer" -> "designed to enable"
... "designed to enable a consistent UX"

luis: maybe something more releaxed
... enabled more relaxed, better

<tlr> "designed to enable a consistent user experience across..."

tlr: happy to do that

luis: second comment
... about website owners operating tls protected sites should anticpate mobile use
... may have constrained capabilities OR restricted trust anchors
... sounds like two options
... but really one is a consequence of the other

yngve: point about certificates, it is possible for a CA that is not embedded in clients to be cross-signed by another CA
... similar to EV

tlr: let's keep that out of our document
... have any concrete proposals?
... my inclination would be to take the text for first comment as agreed above, and taking luis's second change as is currently phrased

RESOLUTION: issue closed, move on

tlr: suggests yngve takes issue to phrase your problem more generally
... and tlr should close out 130

<tlr> ACTION: tlr to change 9.5 in line with ISSUE-130 discussion ago; close issue. [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action02]

<trackbot-ng> Created ACTION-381 - Change 9.5 in line with ISSUE-130 discussion ago; close issue. [on Thomas Roessler - due 2008-02-06].

<tlr> ACTION: yngve to bring up generic techniques for trust root changeover [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action03]

<trackbot-ng> Created ACTION-382 - Bring up generic techniques for trust root changeover [on Yngve Pettersen - due 2008-02-06].

ISSUE-114: Self-signed certificate changeover

ISSUE-114?

<trackbot-ng> ISSUE-114 -- Self-signed certificate changeover -- OPEN

<trackbot-ng> http://www.w3.org/2006/WSC/track/issues/114

tlr: what is currently proposed is a UX that makes cert changeover for SSCs basically impossible
... how do we deal with it
... any way to have changeover to a different SSC is a way to have a MITM attack
... no compelling solutions
... ideas?

<tlr> http://www.w3.org/2006/WSC/drafts/rec/rewrite.html#change-tls-state

yngve: think X509/PKIX had a mechanism for telling what's the next key/whatever
... problem is that SSCs are going to be made by people not aware of it

tlr: doesn't fit use case

yngve: might work for some minor use cases
... not general use case

tlr: can always say that there should be an interaction that provides ample warning and a reasonably complex user interaction that makes it hard to casually accept cert, but we end up having a coin toss between an attack going through and a new legit cert
... we might in fact be best served by saying MAY do override
... tempted to throw out that as a proposal for resolution

tyler: will interested parties be at f2f?

tlr: some, some will be on phone

<bill-d> good drop off

tyler: sort of thing that mez was talking about, fast-track?

tlr: no resolution

RESOLUTION: none

ISSUE-129: Should we say anything about scoring techniques?

<tlr> ISSUE-129?

<trackbot-ng> ISSUE-129 -- Should we say anything about scoring techniques? -- OPEN

<trackbot-ng> http://www.w3.org/2006/WSC/track/issues/129

ISSUE-129?

<trackbot-ng> ISSUE-129 -- Should we say anything about scoring techniques? -- OPEN

<trackbot-ng> http://www.w3.org/2006/WSC/track/issues/129

scribe: quite a bit of recent discussion

tlr: tjh had proposed a rewrite
... not sure he sees a rewrite in there
... wonder if current status is to merge in

<Zakim> ifette, you wanted to jump on this issue

ifette: no
... very against merging language into the document
... feel very strongly against SHOULD
... because a non-binary score is useless to users
... changes in score might be useful as a signal that we should warn something
... but a user seeing some number of bars, a percent, a score, a meter, is totally useless

tlr: will defer this issue
... thanks for having shown up here
... one of the shorter calls in a while, see some of you next week in Palo Alto

Summary of Action Items

[NEW] ACTION: tlr to change 9.5 in line with ISSUE-130 discussion ago; close issue. [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action02]
[NEW] ACTION: tlr to send email to list regarding ISSUE-131 containing full text of new proposal, and will close out the issue [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action01]
[NEW] ACTION: yngve to bring up generic techniques for trust root changeover [recorded in http://www.w3.org/2008/01/30-wsc-minutes.html#action03]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2008/02/13 17:37:00 $