See also: IRC log
<trackbot-ng> Date: 22 January 2008
<hgerlach> yes I am here:-)
<jo> scribe: bryan
<hgerlach> for now or for future as well?
<jo> Close ACTION-608
Francois: I will chair the CT task, so we will close the AI related to finding one.
<trackbot-ng> ACTION-608 Recruit a TF lead for CT closed
Francois: Taskforce homepage needs update to show the change.
Francois: Discussionussion today
covers the CT document updates.
... Cache-control header changes check is still pending. No
reply yet from other W3C groups.
<jo> ACTION-603?
<trackbot-ng> ACTION-603 -- Jo Rabin to find out how to liaise with HTTP NG work -- due 2008-01-03 -- OPEN
<trackbot-ng> http://www.w3.org/2005/MWI/BPWG/Group/track/actions/603
Francois to carry forward the AI for HTTP NG relationship.
Francois: any other alternatives to dependence upon extensions?
Heiko: maybe two step approach
<jo> s/Francoise to/Francois: I will
Jo: discusson with Rhys etc
earlier resulted in decision at the time....
... ... it would be premature to publish something at that
time. .mobi published a style guide draft related to this in
the meantime.
Jo: providing guidance on server config on how to achieve. W3C could do something similar, but we could also note in the document the option for a minimum control based upon headers.
Heiko: we do not need headers for access to mobile OK sites.
Jo: we need them if we want to make the proxy aware of the user agent's basic control preferences.
Heiko: so this does not depend upon knowledge of mobile OK by the user agent
Jo: correct
<jo> [jo explained that the dotMobi paper describes use of Vary and no-transform which are pre-existing headers]
Jo: three possible actions: one,
to note that we can use no-transform without extensions, and
vary header also.
... second, to publish a note that we can do something new
consistent with .mobi style guide
<jo> [3 possibilities are: a) do nothing, b) publish a short note on what could be done with existing headers, basically the same as the dotMobi style guide and c) include a note to this effect in the Guidelines document]
Heiko: could combine with other approaches e.g. via powder and mobile OK, to define how the user knows about mobile OK sites
Jo: there are two aspects: (1)
how to know when to transcode, e.g. based upon heuristics as
Sean mentioned earlier; (2) control aspects
... we can address the control aspects separately
Francois: the control features are listed as to be defined, and we should make the values more precise, recommend something
Jo: the document does need a summary of the proposed changes
Francois: do we want to think about the values on the mailing list?
Jo: we can summarize them and put text in the draft
<jo> ACTION: Jo to add a section summarising the proposed values and proposed names to next draft of Guidelines [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action01]
<trackbot-ng> Created ACTION-622 - Add a section summarising the proposed values and proposed names to next draft of Guidelines [on Jo Rabin - due 2008-01-29].
Francois: next topic, about
requirements part
... re the editor's note, (is it needed); I think the section
is valuable for the reader
Jo: agrees, and proposes to make it a summary of features enabled by this spec; but it's a description then, not requirements
Francois: yet its a description and maybe fits better in another section
Jo: will recast the section as a summary
<jo> ACTION: Recast the Requirements section as a Summary and make it accurate as compared with the details in section 4 [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action02]
<trackbot-ng> Sorry, couldn't find user - Recast
<jo> ACTION: Jo to Recast the Requirements section as a Summary and make it accurate as compared with the details in section 4 [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action03]
<trackbot-ng> Created ACTION-623 - Recast the Requirements section as a Summary and make it accurate as compared with the details in section 4 [on Jo Rabin - due 2008-01-29].
<jo> Close ACTION-604
<trackbot-ng> ACTION-604 Elaborate current draft especially in the areas of requirements closed
Francois: wondering if we can define dangerous content more precisely or we should remove the note
<francois> Bryan: Before the call, I was writing some contents on the list. I think it's useful for us to note there are exceptions to the rules, but not go into details
<jo> Proxy States
<francois> ... We should say clearly that whenever a Proxy does that, it MUST say so
Francois: what about parallels with child protection?
Jo: wanted to say that the proxy
has certain functions that it will perform, and that there are
cases in which it is acceptable if those functions are provided
with active agreement
... agreement of the user and content provider, e.g. removing
images. But we need to say something otherwise there will be
misunderstandings in deployments.
Francois: if we add extensions e.g. except-for, does that help here?
Jo: the statement was written as a general statement, e.g. whether the proxy transformed content may rely upon pre-defined agreements
Heiko: believes content blocking is a separate function and should not affect the controls via no-transform; how is content filtering related to transcoding
Jo: believes filtering is not related
Francois: agrees
... so what is the agreement, should we mention that there are
exceptions?
<jo> [jo used child protection as an analogy to try to justify the proxy intervening in transparent mode, but now regrets using that analogy :-(]
Jo: if a proxy gets no-transform
on its own, it should not do fix-up on its on; if there is no
directive, it could fixup the content
... it should be possible to setup a "session" to operate in
no-transform mode
<hgerlach> sorry I have to leave for the doc, changed to mobile
Jo: the proxy could become passive, e.g. upon no-transform reception, and a user interacting with the proxy instructs the proxy to no transform e.g. with a particular domain
Francois: so there seems to be an
exception to the rule in the section, is that what we
want?
... we could continue this on the list
<jo> ACTION: francois to initiate discuss on the exception wording ref dangerous content [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action04]
<trackbot-ng> Sorry, couldn't find user - francois
<jo> ACTION: fran to initiate discuss on the exception wording ref dangerous content [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action05]
<trackbot-ng> Sorry, couldn't find user - fran
<francois> ACTION: me to initiate discuss on the exception wording ref dangerous content [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action06]
<francois> ACTION: me to initiate discuss on the exception wording ref dangerous content [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action07]
<trackbot-ng> Created ACTION-624 - Initiate discuss on the exception wording ref dangerous content [on Marcos Eguillor Fernandez - due 2008-01-29].
<jo> Close ACTION-624
<trackbot-ng> ACTION-624 Initiate discuss on the exception wording ref dangerous content closed
Heiko: add a topic, can POST responses be modified?
<francois> ACTION: Daoust to initiate discuss on the exception wording ref dangerous content [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action08]
<trackbot-ng> Created ACTION-625 - Initiate discuss on the exception wording ref dangerous content [on François Daoust - due 2008-01-29].
Heiko: believes that only HEAD
and GET responses are covered so far
... in section 3
Jo: requests Heiko raise this on the list
Francois: are non-browser web
applications beyond the scope of the document? How can a
CT-aware proxy know this?
... should transformation occur in these cases?
Jo: Bryan added this in
contribution.
... J2ME clients often use HTTP thru proxies, the question is
how to tell.
<francois> Bryan: I can add some text on that.
<francois> ... I would like to see this as part of the DD repository
<francois> ... the only possibility is to track the user-agent
Bryan: will contribute text on this
<jo> ACTION: Bryan to contribute text on detection of non-browser user agent [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action09]
<trackbot-ng> Created ACTION-626 - Contribute text on detection of non-browser user agent [on Bryan Sullivan - due 2008-01-29].
Francois: how can a CT proxy interact with a non-browser client user? Is the basic assumption that transform should not occur in non-browser environments?
Jo: agrees, our focus should be limited to browsing
Bryan: there are other use cases where CT functions will be useful also
Jo: in browsing use cases, the the origin server needs to be CT-aware
Bryan: agrees, the CT proxy should not break things
<jo> [in order to stop image format manipulation etc.]
<jo> [in non browsing use cases the use cases are even less clear]
<jo> ACTION: Jo to make clear in the scope that we are talking browsing here [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action10]
<trackbot-ng> Created ACTION-627 - Make clear in the scope that we are talking browsing here [on Jo Rabin - due 2008-01-29].
Francois: we could recommend that in a CT-unaware / non-browser case the CT proxy should not get in the way
Francois: would like to see the document move forward; not much more to change before FPWD, e.g. extension values, whether to use them or not...
Jo: preference is to have a draft ASAP in Feb.
Bryan: agrees
Jo: will not be in the next
week's call, and will try to get a new draft out by Friday; the
group should review it, and set a goal for lengthy discussion
in Seoul
... goal for FPWD decision in Seoul
Francois: agrees, will update the roadmap based upon that
<jo> ACTION: Jo to produce draft 1d by Friday [recorded in http://www.w3.org/2008/01/22-bpwg-minutes.html#action11]
<trackbot-ng> Created ACTION-628 - Produce draft 1d by Friday [on Jo Rabin - due 2008-01-29].