See also: IRC log
dom: Normative or informative --
it can be informative if you won't require people to conform to
it. BP's are informative because people use them as a
guide.
... MobileOK is normative, because we want people to conform to
it.
... When BP was chartered it was decided that the CT work would
be informative.
... The charter was not about creating new technologies.
... For normative docs the w3c patent policy applies, which
means members with patents agree to give royalty free
licenses... this does not apply to informative docs.
... We can't make them normative for the time being. The
question is the legal risk more expensive than rechartering the
group to make it normative?
DKA: What about amending the charter at extension time?
dom: If we want to change from
informative to normative, you need to recharter the
group.
... We have to propose something to the w3m, the AC reviews the
charter for four weeks, and then a 45 day grace period for
rejoining the group.
... It takes about a month or two.
... Then republish as WD...
... Then 150 days before published as PR.
<Zakim> Kai, you wanted to ask about a mismatch between the intention of the document and what a normative specification represents
Kai: When you look at a
specification, it's supposed to be something that is
implemented, not a lot of room to interpret, etc. A guideline
has a lot more wiggle room.
... How do we say they've fulfilled the specification?
dom: My reading is that they're
testable.
... WCAG is testable for instance.
DKA: Yes, these should be testable, the work is specific enough.
Kai: In the end you have to say "yes or no" to whether it's adhered to.
<group pauses for station identification>
jo: Propose continuing the group
to end of charter with doc being informative. When the group
recharters I suggest that we take the work continue in a
normative fashion.
... We don't lose much, only 6 months.
dom: We can extend the group without having an AC review for up to a year...
jo: The group probably should be rechartered if it's going to go beyond this year.
<DKA1> PROPOSED RESOLUTION: The CT document should be taken to its conclusion under the current chartered (non-normative) and therefore the question of whether to change it to a normative document should be put off to the future re-chartering.
jo: Can we change an informative rec into a normative rec?
francois: Maybe we can make normative tests that reference the informative guidelines?
jo: We don't want to lose momentum.
DKA: I'm concerned that it will impact the usefulness of the doc in the long run.
jo: Anyone who wants to conform to it will conform to it.
DKA: I don't think it will impact
the conformance but the IP.
... Because it won't be covered under the w3c IP policy.
... The chance of someone having patents on the HTTP stuff is
probably low.
jo: We've had this discussion
already. The rational thing to do is to take legal
advice.
... There are many people who are not w3c members who have
patents and are not obliged to declare them..
Kai: Why is this so important with this document?
DKA: This doc is the one that
gets furthest down the road towards implementing new technology
-- but it is NOT introducing...
... I think we should accept this resolution and then seek
legal advice, give an action to a w3c-er to find out.
dom: Asking the lawyer will result in a no.
jo: Perhaps we can refine the question.
DKA: Is there a way we can
continue with the doc as it is, but work in the background so
we don't lose time before rechartering?
... We get it to CR, and leave it there until July of 2009.
jo: How does this mitigate IP problems?
SeanP: Can we ask for testimonials?
<Zakim> edm, you wanted to ask about publishing the doc as is or inserting some statement re its anticipated normative future (or not...)
SeanP: Wouldn't necessarily be enforceable, but make it feel better.
edm: Can we expose what we're talking about now in the document? Say that it's anticipated to become normative?
DKA: You say keep it informative, but put wording in saying "sorry, we forgot to make it normative?"
edm: Is it better to do it as we are, or to spell it out?
Kai: I don't see the problem. If we make it a normative document, the process deals with it.
jo: The issue is that we have to
change the charter to make it normative. We can't
... produce normative stuff in this charter.
dom: We could start rechartering now...
jo: I would be uncomfortable rechartering without knowing the future of the MWI.
dom: I don't believe the group will close regardless of what happens to the MWI.
DKA: It seems like the work will go on regardless.
SeanP: There's talk about slowing down the momentum if we recharter...
DKA: It will slow things down, the AC has to review it.
dom: Plus there's the delay of getting your AC rep to rejoin the group, etc.
DKA: Plus writing the charter, etc time for the staff.
Kai: There's no short cut?
dom: I don't believe there is
when there's a patent policy implication.
... People still had to disclose patents, even for informative
docs.
Kai: Now if the doc remains informative, this patent stuff isn't an issue...
DKA: Yes, but you could be exposing the entire ecosystem down the line...
jo: But only adding in exposure from the WG members.
dom: People in the WG are
developing this, and don't want to put things in there that are
patented, even by those outside the group.
... If you have patents and are in the group, you are giving a
license.
matt: If this WG is representative of the ecosystem as a whole, and the WG doesn't believe it's patentable, that's a good indicator that it might not be.
DKA: That's why I'm leaning
towards @@
... Perhaps the new charter includes a conformance document
that references this doc... very lightweight.
dom: But you lose the patent
commitments...
... We should take the resolution now or if we make it
normative, start the recharterting process now
... and otherwise keep it informative and leave it alone.
<Zakim> rob, you wanted to note that Openwave las lots of patents, some that might be relevant - but I haven't checked through them to know if they are essential
rob: We've got lots of patents that might be relevant, haven't checked them, haven't gotten the lawyers to look at them either.
DKA: The downside of going into this process is that it might trigger a whole lot of legal searches, etc
dom: It's not a downside.
DKA: Do we think rechartering would significantly delay things?
dom: The biggest risk is that we lose members.
jo: It's not clear that members who signed up through December would be happy to go beyond that.
dom: There's always the chance for participants to leave at any time.
<Zakim> Kai, you wanted to ask about going normative with an unfinished document and the legal issues involved
Kai: I wouldn't bring a doc to the lawyers until it's just about done.
dom: There are various calls for
exclusions during the publishing process.
... You never quite get to see the final document before it's
finished.
DKA: Two options: 1. Restart
rechartering process now, where the only changes: would be the
informative to normative change for this doc, as well as extend
the WG, 6 months into 2009
... Option 2: Do nothing, issue document non-normatively,
possibly revisit in December, but no guarantees.
... Having the document being normative would shield you,
somewhat from member-IP.
dom: Anything they don't exclude
you get a license for.
... Very few charters have been rejected after AC review.
q
matt: If the charter ran out in December, and you wanted a new charter, would there be more work that you might want to add to the WG than just these?
DKA: The talk right now is that
we just want to wind down the work that's in the charter... and
then <mixed metaphor about zombies or something>
... I favor the idea of not adding more work items.
jo: "Option 2: Do nothing, issue document non-normatively, possibly revisit in December, but no guarantees." -- plus get legal advice.
1 vote for option 2, option 1 had five or so votes...
dom: Jo, voting for option 2, is there a strong objection to 1?
jo: I will not object given the strength of feeling on this but there is no chance of my working on writing a new charter given my current work load
DKA: I think people will look
even closer when we recharter at the CT stuff.
... Let's take a resolution that reflects the rough
consensus.
Kai: Option 1 still needs everyone to check their legal stuff before we recharter.
matt: Does this mean jo leaves asap or at end of the year?
PROPOSED RESOLUTION: Restart rechartering process now, where the only changes: would be the informative to normative change for this doc, as well as extend the WG, 6 months into 2009
<DKA1> +1
<Kai> +1
<rob> +1
<dom> I object
<SeanP> +1
<dom> NOT
<edm> +1
<dom> ACTION: francois to start the rechartering process [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action01]
<trackbot> Created ACTION-776 - Start the rechartering process [on François Daoust - due 2008-06-23].
RESOLUTION: Restart rechartering process now, where the only changes: would be the informative to normative change for this doc, as well as extend the WG, 6 months into 2009
<jo> I abstain
<edm> scribe: edm
<scribe> scribenick: edm
dka: would like to get the CT document as close as possible to last call working draft status
<francois> latest draft
dka: suggest a quick walkthrough
francois: emphasis on quick - let's focus on the remaining issues
<DKA1> http://lists.w3.org/Archives/Public/public-bpwg/2008Jun/0049.html
francois: CT guidelines focus on the proxy between the user and the server...
<dom> 13 open issues on content transformation guidelines
francois: ... so that content
provider retains some control over proxy behavior
... CT guidelines are organized as request side and response
side
... restrictions appear mostly on the request side - less on
the response side
... let's review the remaining issues
<francois> Close ACTION-756
<trackbot> ACTION-756 Seek further legal advice on the patent issues around CT closed
dka: let's close the informative vs. normative issue - as per this morning's discussion
<francois> Close ACTION-757
<trackbot> ACTION-757 Seek opinion on CT Patent issue closed
<francois> + Close ISSUE-259
francois: issue-187 (old)
<jo> ISSUE-187?
<trackbot> ISSUE-187 -- Link/Rel to MOK versions of non-MOK resources -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/187
francois: ISSUE-187 Link/Rel to MOK versions of non-MOK resources
jo: this relates to ISSUE-222...
ISSUE-222?
<trackbot> ISSUE-222 -- TAG Finding on Alternative Representations -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/222
francois: link element is the
key
... guidelines specify CT proxy should redirect to handheld
representation...
... ...but the server should indicate the medium for the
intended presentation...
dka: what's the negative of using the link element to link to itself?
jo: how do you determine if the representation from the URI is a link to self
<dom> Media descriptors in HTML 4.01 rec
<dom> "Future versions of HTML may introduce new values and may allow parameterized values. To facilitate the introduction of these extensions, conforming user agents must be able to parse the media attribute value as follows:
<dom> 1. The value is a comma-separated list of entries. For example,
<dom> media="screen, 3d-glasses, print and resolution > 90dpi"
<dom> is mapped to:
<dom> "screen"
<dom> "3d-glasses"
<dom> "print and resolution > 90dpi"
<dom> "
<jo> PROPSOSED RESOLUTION: Link header/element with media type of handheld and an empty URI indicates that the current resource is intended for handheld presentation
jo: we need means for the origin server to signal that this representation is meant for this presentation
dom: points out a difference
between the URI referring to the representation of the resource
and/or the representation
... if you send the right user agent header you should get the
right representation
jo: some URIs point to abstract resource, some to specific representations of this resource
dom: CT guidelines suggest not transforming anything that claims to be mobile friendly - whether or not it really is
<dom> maybe "s/the current resource/the current representation/"
jo: in response to mobile headers, whose mobile friendly content should apply - origin server's or proxy's?
<jo> PROPOSED RESOLUTION: Having requested a resource with mobile headers, a Link header/element with media type of handheld and an empty URI indicates that the current representation of the resource is intended for handheld presentation
<dom> +1
dom: the decision of what is appropriate representation rests with the content provider
<dom> Results of testing 406 responses on mobile browsers
Francois: 404 response works, 406 not necessarily for all browsers
dka: media handheld is what is being supported now, other media types could be considered when get introduced
<Zakim> jo, you wanted to say something meaningful
<dom> profile attribute on HEAD in HTML
<jo> PROPOSED RESOLUTION: a Link header/element with media type of handheld and a URI referring to a location within the resource indicates that the current representation of the resource is intended for handheld presentation
dka: would this allow to close ISSUE-222?
<rob> +1
<francois> +1
<jo> +1
<Kai> +1
<DKA1> +1
<SeanP> +1
RESOLUTION: a Link header/element with media type of handheld and a URI referring to a location within the resource indicates that the current representation of the resource is intended for handheld presentation
<Kai> ScribeNick: Kai
DKA: only the tag thing needs to be resolved for this issue
<jo> PROPOSED RESOLUTION: Make the point explicitly in the document that an empty href on a link header/element means that this reource is capable of being rendered in a way that is suitable for handhelp presentation, but that without the above link element there is a question as to whether this representation is or is not suitable for handheld
Jo: if you have an emtpy href it could mean it is or could be rendered as handheld
francois: the important part is the fragment. you could link to yourself
dom: does that effect the resolution we just took?
francois: no not really.
<jo> PROPOSED RESOLUTION: Make the point explicitly in the document that an href which refers to this resource but which does not have a fragment identifier on a link header/element means that this reource is capable of being rendered in a way that is suitable for handheld presentation, but that without the above link element there is a question as to whether this representation is or is not...
<jo> ...suitable for handheld
<francois> +1
<DKA1> +1
RESOLUTION: Make the point explicitly in the document that an href which refers to this resource but which does not have a fragment identifier on a link header/element means that this reource is capable of being rendered in a way that is suitable for handheld presentation, but that without the above link element there is a question as to whether this representation is or is not...
<SeanP> +1
<dom> ISSUE-261?
<trackbot> ISSUE-261 -- Scoping 200 responses that should be regarded as 406 responses -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/261
francois: Topic ISSUE-261
<edm> ISSUE-261?
<trackbot> ISSUE-261 -- Scoping 200 responses that should be regarded as 406 responses -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/261
francois: I am not sure how many
websites will send human readble responses about an error
... Aaron was going to send some stats on this
... he has not yet, but could bring stats within 2 weeks
... is our recommendation useful and needed?
dom: does dotmobi have studies on this?
jo: not that I know of
Dka: is that a meaningful question? If one or two high use sites do it, that would already be as damaging as many low use sites
francois: if only a few are doing that, proxies could list those and send different headers
jo: however they do it does not concern the guidelines#
dka: the guidelines are written
thinking that more and more site adapt, operators use
proxies...problems will become bigger problems, which is why we
are setting these rules down.
... this is the whole point...anticpation
... we try to codfiy it before it becomes a problem
francois: we already have enough restritction on http header modification
dka: key is are recommeding this
approach or not?
... for when a 200 response is sent but there is no link
header. If they do see it they should consider it.
... we are justified to put this in the doc
SeanP: so you say we will have to do something about this anyway?
DKA: yes
jo: we could not mention bogus
responses
... or we say a 200 response like that is a rejection
... if there is no no-transform this is intentional
SeanP: I think we should mention it. we don't want to be too vague.
DKA: we should be explicit
<jo> http://lists.w3.org/Archives/Public/public-bpwg-ct/2008Jun/0039.html
jo is going through his email
jo: proposes to use the pertinent part of the document, with some editorial changes
<jo> jo quotes the following:
<jo> The logic of 4.1.2 would then be:
<jo> Request resource with original headers
<jo> - If the response is a 406 response, re-request with altered headers
<jo> (unless the 406 response contains no-transform).
<jo> - If the response is a 200 response,
<jo> -- if the response contains vary: User-Agent, an appropriate link
<jo> element or header, or no-transform, forward it
<jo> -- otherwise assess (by unspecified means) whether the 200 response is a
<jo> bogus one
<jo> --- if it is not, forward it
<jo> --- if it is, re-request with altered headers
jo: we would not want this to become a normative algorythm
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-261, adopt Jo's proposal for a process for a process for CT bogus 200 response detection (4.1.2) and close ISSUE-261.
<jo> +1
<francois> +1
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-261, adopt Jo's proposal for a process for a process for CT bogus 200 response detection (4.1.2) and close ISSUE-261.
<dom> +1
RESOLUTION: Regarding ISSUE-261, adopt Jo's proposal for a process for a process for CT bogus 200 response detection (4.1.2) and close ISSUE-261.
<francois> Close ACTION-673
<trackbot> ACTION-673 See if he can get some figures that scope the problem of bogus 200 responses closed
<francois> Close ACTION-768
<trackbot> ACTION-768 Ping Aaron on scoping bogus 200 responses closed
francois: so I don't have to bother aaron
<francois> Close ACTION-771
<trackbot> ACTION-771 Send a proposal on using meta data to alert CT-proxy that this is a rejected response even if it's a 200 closed
<jo> ACTION: Jo to edit 4.1.2 according to above resolution [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action02]
<trackbot> Created ACTION-777 - Edit 4.1.2 according to above resolution [on Jo Rabin - due 2008-06-23].
francois: ISSUE-257?
ISSUE-257?
<dom> ISSUE-257?
<trackbot> ISSUE-257 -- Clarification of section 4.1.2 -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/257
francois: I would like to resolve to rename the section
jo: i would prefer if we address
specific points
... (going through document)
... Section 4.1.2
<DKA1> a) the user would be prohibited from accessing content as a result of a
<DKA1> 406 or bogus 200 response;
<DKA1> b) the user has specifically requested a restructured desktop experience;
<DKA1> c) the request is part of a session of some sort and either it is
<DKA1> technically infeasible not to adjust the request because of earlier
<DKA1> interaction, or because doing so preserves a consistency of user experience.
<DKA1> In that section my understanding of what we are trying to do is
<DKA1> identify that the request headers should not be altered unless:
<DKA1> a) the user would be prohibited from accessing content as a result of a
<DKA1> 406 or bogus 200 response;
<DKA1> b) the user has specifically requested a restructured desktop experience;
<DKA1> c) the request is part of a session of some sort and either it is
<DKA1> technically infeasible not to adjust the request because of earlier
<DKA1> interaction, or because doing so preserves a consistency of user experience.
jo: i think together with renaming this is the essence of what we are trying to say
francois: we still need to deal with user preferences
jo: yes, need to have a
discussion on it
... so we would remove the unnumbered bullets and reword 1
through 3
... do we need very specific text here?
... there are three cases to consider
dka: that removes the discussion of prior knowledge, server behavior etc.
jo: yes it does
dka: it removes the ability for CT proxies to rely on user preferences
francois: that is a different discussion
dka: the request doesn't have to be at that point
jo: that's for us to define
seanp: can look at the exact
text?
... tick...tick...tick...tick
<jo> Proxies should not intervene in methods other than GET, POST, HEAD and PUT.
<jo> If there is a no-transform directive present in the request the proxy should not modify the request headers.
<jo> The proxy should not modify request headers unless:
<jo> a) the user would be prohibited from accessing content as a result of a
<jo> 406 or bogus 200 response;
<jo> b) the user has specifically requested a restructured desktop experience;
<jo> c) the request is part of a session of some sort and either it is
<jo> technically infeasible not to adjust the request because of earlier
<jo> interaction, or because doing so preserves a consistency of user experience.
<jo> Note:
<jo> Rejection of a request by a server is taken to mean both a HTTP 406 Status as well as HTTP 200 Status, with content indicating that the request cannot be handled - e.g. "Your browser is not supported"
dka: this replaces all up to point 3) in 4.1.2
francois: can we change the name of the section?
SeanP: this is to remove the duplication and make it clearer?
<DKA1> PROPOSED RESOLUTION: Adopt Jo's wording for 4.1.2 of the CT document (pending some editorial "tweaking") and close ISSUE-257.
dka: any objections?
<DKA1> +1
<francois> +1
RESOLUTION: Adopt Jo's wording for 4.1.2 of the CT document (pending some editorial "tweaking") and close ISSUE-257.
francois: could we agree on
issue-243 to leave it and address it later?
... it's linked to POWDER and can probably wait?
<jo> ACTION: Jo to add the stuff on possible use of OPTIONS to the appendix [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action03]
<trackbot> Created ACTION-778 - Add the stuff on possible use of OPTIONS to the appendix [on Jo Rabin - due 2008-06-23].
francois: I will take the action
<dom> ACTION-777?
<trackbot> ACTION-777 -- Jo Rabin to edit 4.1.2 according to above resolution -- due 2008-06-23 -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/actions/777
<DKA1> PROPOSED RESOLUTION: Close ISSUE-243.
<francois> PROPOSED RESOLUTION: Close ISSUE-243 and mention it in "Scope for Future Work" appendix
<DKA1> PROPOSED RESOLUTION: Close ISSUE-243 and mention this topic in the appendix as scope for future work.
<francois> +1
RESOLUTION: Close ISSUE-243 and mention this topic in the appendix as scope for future work.
ISSUE-224?
<trackbot> ISSUE-224 -- How should we make an assessment of the impact of the various possible guidelines on real web sites -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/224
francois: don't know how to address that?
jo: it is not so much of an issue today.
dom: the cr phase would be for that
<DKA1> PROPOSED RESOLUTION: regarding ISSUE-224, we are no longer as worried about the impact of these guidelines on existing content, so this issue can be closed and the topic left to the exit-criteria from CR (testing implementations).
<francois> +1
<DKA1> PROPOSED RESOLUTION: regarding ISSUE-224, we are no longer as worried about the impact of these guidelines on existing content, so this issue can be closed and the topic left to the exit-criteria from CR (testing implementations) - and close ISSUE-224.
<francois> Close ACTION-774
<trackbot> ACTION-774 Check the OPTIONS method for ISSUE-243 closed
<jo> +1
RESOLUTION: regarding ISSUE-224, we are no longer as worried about the impact of these guidelines on existing content, so this issue can be closed and the topic left to the exit-criteria from CR (testing implementations) - and close ISSUE-224.
<DKA1> lunch!
dka: we will continue to go through issues after lunch. Will do scheme discussion after second break.
<DKA1> Returning from lunch.
<rob> Scribe: rob
<scribe> scribenick: rob
dka: 6 issues remain...
<dom> ISSUE-223?
<trackbot> ISSUE-223 -- Various Items to Consider for the CT Guidelines -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/223
francois: 1-6 already dealt
with
... point 7 seems like nothing we can do without inventing new
technology
... POWDER could be used in future
jo: points 7,8,9,11 are the same - future work
<DKA1> PROPOSED RESOLUTION: From Jo's list (ISSUE-223) we will consider points 7, 8, 9 and 11 as out of scope for the current work (but possibly in scope for future work).
<francois> +1
<SeanP> +1
+1
RESOLUTION: From Jo's list (ISSUE-223) we will consider points 7, 8, 9 and 11 as out of scope for the current work (but possibly in scope for future work).
<jo> ACTION: Jo to transcribe points 7 8 9 and 11 of ISSUE-223 into Scope for future work [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action04]
<trackbot> Created ACTION-779 - Transcribe points 7 8 9 and 11 of ISSUE-223 into Scope for future work [on Jo Rabin - due 2008-06-23].
francois: point 10 is MobileOK required for CT-proxies?
dka: seems obvious, it's overloading the doc
jo: in doc sec 4.4 we said CT-proxy output should validate according to a formal standard
<DKA1> PROPOSED RESOLUTION: On point 10 of ISSUE-223 (whether CT proxies should be MobileOK) we should remain silent.
<SeanP> +1
<DKA1> +1
+1
<Kai> +1
RESOLUTION: On point 10 of ISSUE-223 (whether CT proxies should be MobileOK) we should remain silent.
francois: point 12 - if content is mobileOK then can it be transformed?
<DKA1> PROPOSED RESOLUTION: On point 12 of ISSUE-223 (whether CT proxies should refrain from transforming MobikeOK content) we should allow for the possibility that CT proxies should be able to transform MobikeOK content but that they should take into account MobileOK-ness as part of the heuristics involved in determining whether content is mobile-friendly.
<SeanP> +1
kai: content could still be transformed (from mobileOK to mobileOK) because a CT-proxy knows somehting about the device that the original content didn't account for
jo: RFC2616 has an example about medical imaging content that mustn't be altered
<DKA1> PROPOSED RESOLUTION: On point 12 of ISSUE-223 (whether CT proxies should refrain from transforming MobikeOK content) we should allow for the possibility that CT proxies should be able to transform MobikeOK content but that they should take into account MobileOK-ness as part of the heuristics involved in determining whether content is mobile-friendly (but remain silent on how you check if something is mobileOK).
<francois> +1
<SeanP> +1
+1
<Kai> +1
<dom> +1
<DKA1> PROPOSED RESOLUTION: On point 12 of ISSUE-223 (whether CT proxies should refrain from transforming MobikeOK content) we should allow for the possibility that CT proxies should be able to transform MobileOK content but that they should take into account MobileOK-ness as part of the heuristics involved in determining whether content is mobile-friendly (but remain silent on how you check if something is mobileOK).
RESOLUTION: On point 12 of ISSUE-223 (whether CT proxies should refrain from transforming MobileOK content) we should allow for the possibility that CT proxies should be able to transform MobileOK content but that they should take into account MobileOK-ness as part of the heuristics involved in determining whether content is mobile-friendly (but remain silent on how you check if something is mobileOK).
<jo> +1 (in respect of Mobikes)
<jo> ACTION: Jo to add text to section 4.4 referencing above resolution on mobikeOK [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action05]
<trackbot> Created ACTION-780 - Add text to section 4.4 referencing above resolution on mobikeOK [on Jo Rabin - due 2008-06-23].
ISSUE-241?
<trackbot> ISSUE-241 -- Tokenization of URIs -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/241
francois: do we need to say anything about changing links in an adapted session to "made-up links" to the CT-proxy's own domain so it can invent new URLs for sub-pages, javascript events, etc
dom: so what about rewriting URLs in a page of search-results?
seanP: they all get rewritten
jo: if you follow what it says in sec 4.1 you will then fall out of the transforming loop
dom: URL re-writing does overwrite the Host: request header
jo: it would be inconsistent to allow users to go to the operator portal and the operator to rewrite URLs to the whole of the Internet to then assume they are compliant
<dom> [I would request at least that the issue be re-titled]
francois: need to ensure content-providers at least can see the requests as faithfully as possible when appropriate
jo: this touches on the undefined concept of a "session" being entered where the CT-proxy is unavoidably explicitly addressed by all requests from the handset
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request to a new server, this is a different "session" and therefore the content must be tasted.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request to a new "web site", this is a different "session" and therefore the content must be tasted (allowing for different heuristics for determining whether request is to a different or the same "web site.")
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request to a new "web site", this is a different "session" and therefore the content must be tasted (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example.)
jo: key issue is if CT-proxy is transforming content for site A and then makes bad decisions in continuing transformation following links to site B that would normally be mobile-friendly
kai: domain name isn't a great distinguishing feature for websites, eg images often served from different domains
seanP: in the CT world, we tend to think of a "session" as something between the handset and CT-proxy, not between bandset and website
dom: currently guidelines allow you to carry on transforming all the time (even after following a link to mobile-friendly site) after you have started transforming a website that needed transformation
jo: do we need to distinguish
between "linked resources" (eg <a href="...">) and
"included resources" (eg <img src="..." />)
... ?
dka: do we stand a chance of
resolving this issue today? Or should we leave it for future
work?
... should we tease apart the two sides of the CT-proxy and
define "sessions" on both sides?
... or concentrate on rules for "tasting content"?
jo: both in my opinion
<DKA1> PROPOSED RESOLUTION: Except in the case of https, the content transformation guidelines document does not differentiate between URI rewriting and so-called "transparent" proxy operation.
<DKA1> PROPOSED RESOLUTION: For CT guidlines, for included resources within a page, content tasting is not required.
francois: for included resources tasting is clearly not required. It's only be for linked resources that the question of tasting content arises
+1
<DKA1> PROPOSED RESOLUTION: For CT guidelines, for included resources within a page, additional content tasting is not required.
<Kai> +1
<francois> +1
<SeanP> +1
RESOLUTION: For CT guidelines, for included resources within a page, additional content tasting is not required.
<Kai> ScribeNick: Kai
jo: should you then taste all
linked resources?
... this is where get into the end to end session
... it would be simpler if we said all linked resources must be
tasted
<DKA1> PROPOSED RESOLUTION: For CT guidelines, except in the case of https, the content transformation guidelines document does not differentiate between URI rewriting and so-called "transparent" proxy operation.
rob: tasting linked resources may end in superflous tasting
jo: a linked resource is the
target of an a element...for sake of simplicity
... following a hyper link is a linked resource...what would
that mean?
SeanP: might mean extra requests
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linkd resource to a new "web site", content must be tasted (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example.)
rob: perhaps you should abide by the same tasting rules if you don't know the resource
jo: what does not know
mean?`
... tasted before cache expiry level
DKA: can we focus on resolutions?
RESOLUTION: For CT guidelines, except in the case of https, the content transformation guidelines document does not differentiate between URI rewriting and so-called "transparent" proxy operation.
jo: for the second resolution..any linked resource should be tasted with respect to cache expiry
dka: I don't think so. this
brings up all the issues of tasting...am I putting two bids on
an auction, putting two books in my basket, etc.
... it's a non starter
jo: we are exploring what
ifs
... never taste or always taste linked resources
francois: dan's point relates to
the HTTP method. If there is a form it should not trigger a new
tasting.
... can this be explained with GET?
dka: I don't think it is possible.
SeanP: another issue is whether a
user wants a desktop experience
... then you wouldn't need to taste it
francois: that is a separate issue
SeanP: the resolution says that it must be tasted.
rob: it should be what is says 4.1
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example.)
jo: if we say that linked resources are a different case then we always have content tasting. On the other hand never do that, which it won't be and I don't know why it is not always
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example.)
rob: it is not clear to me either considering 4.1
francois: if there is a session started, if there is another tasting for the same session......
dka: jo you are saying we don't need the thing about new website, because it is dealt with?`
jo: no, not exactly.
dka: so what is wrong with the
proposed resolution?
... (reading the resolution)
jo: rob is thinking why not always taste, but prior knowledge has been taken out....how does that change what you said?
rob: dan is talking about a resource to a new website, using the domain name as a definition....it is probably as good as taking out all prior knowledge
jo: the prior knowledge relates
to black and white list discussion
... the problem there is if the server alters its behavior it
is difficult to change lists. It would be better to do this
dynamically.
dka: I don't see where we have white and black lists
jo: we had this last week
... if we can do this dynamically, without ref to prior
knowledge or blacka nd white lists it sets an expectation that
server behavior is taken account of
... question about the sesssion is related to the prior
knowledge...
... so this is like lists, you could make equal
assumptions
... linked resources...how often do you need to taste...would
also ensure you are fresh
dka: we are not introducing wording that would have a bad effect.
jo: but we need to take session
out
... this is why we are teasing this apart
dka: so hence my resolution
jo: but it may be more
granular
... if you have a domain with people having differnet websites
underneath same domain...for example
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that the proxy should not taste every linked resource.
SeanP: I think I agree with Dans resolutio before..
rob: we should say do taste if you go to a different website, but not define it as such
jo: let's take both of these
resolutions then
... section 4.3...we are saying that you have to re-request, if
headers have been modified
... this means if the heuristics are wrong you may be wrong
francois: there are other ways to control transformation.....
jo: I understand but lets take these resolutions, after modification to remove the word session
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example. And remove the word "session" from the previous resolution on 4.1.2.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that the proxy should not taste every linked resource.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we agree that the proxy does not have to taste every linked resource.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we agree that the proxy does not have to taste every linked resource (for the sake of clarity).
RESOLUTION: Regarding ISSUE-241, in the CT guidelines we agree that the proxy does not have to taste every linked resource (for the sake of clarity).
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example. And remove the word "session" from the previous resolution on 4.1.2.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example.) And remove the word "session" from the previous resolution on 4.1.2.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example. And remove the word "session" from the previous resolution on 4.1.2. And close ISSUE-241. And subsidize soybeans
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example. And remove the word "session" from the previous resolution on 4.1.2. And close ISSUE-241.
RESOLUTION: Regarding ISSUE-241, in the CT guidelines we should state that when the proxy makes a request for a linked resource to a new "web site", then what's in section 4.1.2 should happen (allowing for different heuristics for determining whether request is to a different or the same "web site" but giving "hitting a new domain name" as a good example. And remove the word "session" from the previous resolution on 4.1.2. And close ISSUE-241.
<dom> ISSUE-255?
<trackbot> ISSUE-255 -- Subdomain and Path as a heuristic in content transformation -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/255
<SeanP> scribe: SeanP
<scribe> scribenick: SeanP
<jo> [jo makes note to self ref Linked Resources and Included Resources plus Form Submission]
<dom> PROPOSED RESOLUTION: re. ISSUE-255, drop mention of examination of URI from 4.1.2 as it's a heuristic to scope rejected 200 responses, detailed in 4.4, and 4.1.2 already precises to examine the response (with a link to 4.4)
<francois> +1
<jo> ACTION: Jo to enact changes sugegsted by the previous 4 resolutions [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action06]
<trackbot> Created ACTION-781 - Enact changes sugegsted by the previous 4 resolutions [on Jo Rabin - due 2008-06-23].
RESOLUTION: re. ISSUE-255, drop mention of examination of URI from 4.1.2 as it's a heuristic to scope rejected 200 responses, detailed in 4.4, and 4.1.2 already precises to examine the response (with a link to 4.4)
Francois: Do we still want to do what's in 258?
Jo: Not sure requirements belong here, but some people like them.
Francois: Used to find this useful, now that I've been working on it a while, think it should be removed.
DKA: If you are new to the W3C,
then you might need it. Could leave it in for historical
reasons.
... Could move requirements into scope section.
Jo: Requirements are partially wrong.
Francois: Should move them and refresh them.
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-258, move section 3.1 of CT Guidelines document into Scope section (and refresh to synchronize with the rest of the document).
Jo: In 3.1, 1a and 1b are
fine.
... 2a is OK, 2b is not OK, 2c is OK
... 3a is OK, 3b is OK
... 3b needs to be replaced by "is not permissible"
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-258, move section 3.1 of CT Guidelines document into Scope section (and refresh to synchronize with the rest of the document) moving removed requirements into scope for future work.
Jo: 3c is OK
... 3d is OK
... 4a is OK, 5a is not OK, 5b is OK
... 2d is OK
<DKA1> PROPOSED RESOLUTION: Regarding ISSUE-258, move section 3.1 of CT Guidelines document into Scope section (and refresh to synchronize with the rest of the document) moving removed requirements into scope for future work (and close ISSUE-258).
Jo: This could be folded into 3.2 (1, 2, 3, and 4)
RESOLUTION: Regarding ISSUE-258, move section 3.1 of CT Guidelines document into Scope section (and refresh to synchronize with the rest of the document) moving removed requirements into scope for future work (and close ISSUE-258).
Francois: Difference between
regular CT proxy and proxy client combination (like Opera
Mini). Treat proxy/client combo as black box.
... Jo noted that black boxes tend to leak.
... So guidelines should partially apply. Example of where
guidelines should apply is to fix problem where all Opera Mini
requests seem to come from Norway.
Rob: with google proxy, all requests come from the same place too
Francois: but in that case, the Google proxy is a regular CT proxy
<jo> ACTION: Jo to draft text on which aspects of the CT guidelines should be followed by e.g. Opera Mini [recorded in http://www.w3.org/2008/06/16-bpwg-minutes.html#action07]
<trackbot> Created ACTION-782 - Draft text on which aspects of the CT guidelines should be followed by e.g. Opera Mini [on Jo Rabin - due 2008-06-23].
Francois: We should put something about proxy/client combo following guidelines into Appendix.
<francois> Close ACTION-678
<trackbot> ACTION-678 Raise and issue on the distinction between a CT proxy and (say) Opera Mini closed
<dom> ScribeNick: dom
ISSUE-242?
<trackbot> ISSUE-242 -- User expression of persistent and session preferences -- OPEN
<trackbot> http://www.w3.org/2005/MWI/BPWG/Group/track/issues/242
Francois: 2 questions here
... the distinction between persistent and
semi-persistent
... we decided to avoid talking about semi-persistent
settings
... we still have to explain what persistent means
... The problem here is about user preferences and
administrative arrangements
... User preferences could just mean that the operator got the
user to sign an agreement that completely removes control from
the user
... this opens a big hole in the guidelines
... we need to rewrite 3.2 to be explicit (or clear we're not
explicit)
DKA: this about the relation between the user and the CT proxy, right?
FD: yes, unrelated to a specific
request made by the user
... It could be controlled by terms and conditions set by the
operator
EdM: how persistent is persistent? until explicitly changed by the user?
Jo: let's put it another
way
... in the rewritten 4.1.2, the 3 conditions for a proxy to
change the headers: technically required, according by previous
interactions with the server, because the user has specifically
requested a desktop presentation
... for this third bit, what does it mean?
... * for this server specifically
... * or for any request?
... I think that's what's at the heart of this
... I think it's reasonable that the user should be able to say
"for this site, always give me the desktop presentation"
... (if the proxy offers that feature)
<jo> PROPOSED RESOLUTION: It is permissible for the proxy to offer the user a restructured desktop presentation on a 'site' by 'site' basis
EdM: when you say site, is "m.example.com" and "www.example.com" the same site?
FD: we leave this to CT vendors to define
Jo: one of the questions is whether we can say something is permissible if it isn't under law?
DKA: think we're covered by the document license
RESOLUTION: It is permissible for the proxy to offer the user a restructured desktop presentation on a 'site' by 'site' basis
Jo: in 3.1 of the guidelines, we
say that the best place to implement user choice of
presentation is at the origin server
... which comes first? presentation switch on the origin server
or in the CT proxy?
... so it's probably worth noting that it would be useful to
develop that point in BP2
ISSUE: Discuss the option to offer choices of presentation as a best practices for mobile web apps
<trackbot> Created ISSUE-262 - Discuss the option to offer choices of presentation as a best practices for mobile web apps ; please complete additional details at http://www.w3.org/2005/MWI/BPWG/Group/track/issues/262/edit .
Jo: we want to encourage the creation of mobile-friendly content on the origin server, so we probably don't want to promote a practice that would go against that policy
SeanP: so you're saying that a CT proxy should have a persistent preference to always get the desktop version?
Jo: the complication is that if
the user wants a mobile version, and the CT still sends the
desktop headers, the user doesn't get what she wants
... it raises the question of blanket user preferences
SeanP: but that's what most CT do
Jo: but is this the expression of a specific user's choice?
SeanP: I think in reality is that the operator is making the choice
Jo: but is it appropriate? esp.
with the mission of the BP in mind
... a concern is that if all vendors default to
desktop-presentation, this would still be conformant with the
guidelines without helping what we want to achieve
... so my conclusion that a blanket agreement doesn't
constitute a proper expression of user's choice
Francois: my concern is, if the user really wants the default desktop-experience, we would forbid him to have this expressed or recorded?
Jo: it comes back to a matter of
principle
... the CP choice must prevail, unless the user specifically
expresses it
<DKA1> PROPOSED RESOLUTION: If both a mobile and a desktop experience are provided by the origin server, the default should be to provide the user with the mobile experience unless the proxy determines that providing the mobile experience will result in a non-functional user experience.
SeanP: the way of the CT work now is that they send the desktop headers
<Zakim> francois, you wanted to talk about power users
francois: it's probably true that
this would affect very few users (referring to dom's not
minuted comment)
... it's no big deal if only a few users can't express their
preferences if most benefit from it
Jo: I'm not sure we should make assumptions about the existing balance - given how rapidly these consideration change
<Zakim> Kai, you wanted to point out that we need to consider editorial issues when it comes to thinking about content providers views
Kai: editorially speaking, it's sometimes impossible to have a single version adapted to mobile - and creating many versions is expensive
<jo> PROPOSED RESOLUTION: Blanket expression of user preference for restructured desktop presentation is not consistent with encouraging content providers to build specific mobile user experiences and providing a choice of experience at the origin server
Kai: automation is cheap - so leaving the burden on transformation proxy makes economical sense
Dom: but there is a different between content transformation proxies and content adaptation engine used by the CP
Jo: we should note that distinction in the doc
<jo> PROPOSED RESOLUTION: Insert into the document a scoping statement that says that a CT proxy is a CT proxy only if it is not under the control of a content provider. One that is in control of the contrnet provider is an adaptation solution and is out of scope
<jo> PROPOSED RESOLUTION: Insert into the document a scoping statement that says that a CT proxy is a CT proxy only if it is not under the control of a content provider. One that is in control of the contrnet provider is an adaptation solution and is considered to be part of the origin server and is therefore out of scope
<jo> PROPOSED RESOLUTION: Insert into the document a scoping statement that says that a CT proxy is a CT proxy only if it is not under the control of a content provider. One that is in control of the content provider is an adaptation solution and is considered to be part of the origin server and is therefore out of scope
DKA: Vodafone does both - isn't that a bit wishy-washy?
<jo> PROPOSED RESOLUTION: Insert into the document a scoping statement that says that a proxy is a CT proxy only where no prior arrangement exists between the operator of the proxy and a content provider. One where an arrangement exists with the content provider is an adaptation solution, is considered to be part of the origin server and is therefore out of scope
<DKA1> +1ish
<rob> +1
<SeanP> +1
<jo> PROPOSED RESOLUTION: Insert into the document a scoping statement that says that a proxy is a CT proxy in scope of this document only where no prior arrangement exists between the operator of the proxy and a content provider. One where an arrangement exists with the content provider is an adaptation solution, is considered to be part of the origin server and is therefore out of scope
RESOLUTION: Insert into the document a scoping statement that says that a proxy is a CT proxy in scope of this document only where no prior arrangement exists between the operator of the proxy and a content provider. One where an arrangement exists with the content provider is an adaptation solution, is considered to be part of the origin server and is therefore out of scope
PROPOSED RESOLUTION: Blanket expression of user preference for restructured desktop presentation is not consistent with encouraging content providers to build specific mobile user experiences and providing a choice of experience at the origin server
<jo> PROPOSED RESOLUTION: Blanket expression of user preference for restructured desktop presentation is not consistent with encouraging content providers to build specific mobile user experiences and providing a choice of experience at the origin server
+1
<jo> +1
<SeanP> 0
<DKA1> +1
<adam> +1
PROPOSED RESOLUTION: Blanket expression of user preference for restructured desktop presentation is not consistent with encouraging content providers to build specific mobile user experiences and providing a choice of experience at the origin server
Jo: a possible simple fix would to insert simply a note in the document that states this
francois: and remove the part that allows to always request the desktop version
jo: right
<jo> (and remove section 3.2.3)
jo: otherwise, we can require
this to be made on a case by case basis
... a further choice would be to say that when a CP provides
the choice, this should override the blanket agreement
Kai: why this would override the choice of the user?
Jo: the choice the user has expressed is with the CT, not with the CP
SeanP: I would disagree with such a proposal - we have plenty of requests for this
Jo: but the user doesn't really care whether the desktop version comes from the CT or the CP
DKA: we need to consider the
point of encouraging content providers to create mobile
friendly content
... these rules need to be consistent with that overall set of
goals
... also, this task force was kicked off in reaction to the
call from content providers to have more control and visibility
in the proxy layer
... so I think we should be erring on the side of what Jo
proposed
<Zakim> rob, you wanted to say that site-by-site opt-in to restructuring desktop editions could also prevent thos sites from subsequently getting a mobile-friendly edition to the user
Jo: we need to reconcile the need from the users (e.g. always wanting the desktop version), from the needs of CP to control what is served
Rob: site-by-site opt-in to
restructuring desktop editions could also prevent those sites
from subsequently getting a mobile-friendly edition to the
user
... I think default is less an issue than the fact that you can
change your mind
DKA: but most users just don't
know anything about this
... so we're talking about making decisions for users
... with a blanket preference, as a new user, I wouldn't ever
know that there may be a mobile version of a given site
SeanP: there would probably be a link, though
DKA: (which is what Jo raised as an issue for BP2)
SeanP: if you disallow blanket preferences for Desktop-version, you run the risk of people not following the guidelines
Dom: what about requiring a CT proxy that relies on a user preference to always get the desktop version to have a link to get the mobile version served by the site?
Kai: what about the other way around?
DKA: or more generally speaking,
when a CT proxy does transformation, requiring that it puts a
link to the "other mode" of the site?
... In vodafone, lots of the discussions about the default
depends on the device the user has
<jo> PROPOSED RESOLUTION: Put a pin in this discussion noting that the questions that remain to be resolved are:
<jo> a. how the CT proxy should react against specific user preferences when a site becomes more capable
<jo> b. that the default experience should be for the mobile experience
<jo> c. that blanket expression of preferences should be interpreted in the context that if an origin server can provide a choice of experience then it should do so
<jo> d. that CT proxies should, in restructured content, proivide links to alternative representations
Jo: I agree with Rob's point that the user should be alerted when a site changes and becomes mobile-friendly if she had expressed a preference for that site before
<jo> e. how would a CP indicate to a CT Proxy that it offers user choice of representation?
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for desktop presentation and multiple representations are found to exist then the CT proxy server SHOULD prompt the user to inform them of this and ask the user which representation they want to view.
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for desktop presentation and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change the representation.
<jo> PROPOSED RESOLUTION: if there is a blanket user preference asserted for desktop presentation and multiple representations are found to exist then the CT proxy server SHOULD make the user aware of it and provide the user with the option to change the represntataion
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for any specific presentation option and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change the representation.
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for any specific representation option and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change the representation.
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for any specific representation option and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change to one of the alternative representations.
Adam: this seems to be making the
document much more complex
... I wonder if we shouldn't stay silent on this
<DKA1> PROPOSED RESOLUTION: We remain silent on the issue of what explicit prior user consent consists of.
Jo: the key question is the level of user expression consent
Dom: agree; once the user has
sufficiently agreed to it, the ct proxy becomes part of the
user-agent
... (e.g. a user could use a user-defined proxy to do the
transformations he deems necessary)
[discussions on no-transform]
SeanP: one problem with
no-transform is that ct proxies still want to add
headers/footers or rewrite links
... even if the content is set to no-transform
Dom: I think that whether this is
going to be implemented or not is a different discussion on
whether it should be or not
... maybe not all vendors will apply this, but that doesn't
change the fact that we might think this is a good thing
DKA: as an operator, we may actually want to have these changes as a longer term goal
<DKA1> PROPOSED RESOLUTION: if there is a blanket user preference asserted for any specific representation option and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change to one of the alternative representations.
RESOLUTION: if there is a blanket user preference asserted for any specific representation option and multiple representations are found to exist then the CT proxy server SHOULD inform the user of this fact and provide the user with an option to change to one of the alternative representations.
DKA: I think we need to strengthen the clauses around "no-transform"
Rob: when rewriting urls, if one of the links link to a no-transform site, what do you do?
dom: you could redirect there instead of rewriting?
<jo> PROPOSED RESOLUTION: For no-transform responses and for https change text referring to "express user choice" to mean "case by case choice"
<DKA1> PROPOSED RESOLUTION: In the context of tasting content, if header comes back as cache-control no-transform, then CT proxies SHOULD change to transparent proxy operation.
<DKA1> PROPOSED RESOLUTION: I
<DKA1> PROPOSED RESOLUTION: remain silent on all other issues
<francois> ScribeNick: dom
<francois> ScribeNick: francois
DKA: can we exclude visible header/footers from our discussion?
SeanP: Well, the other issue is URI rewriting
DKA: I don't think that's a problem.
<DKA1> PROPOSED RESOLUTION: no means no
DKA: If a site is precisely tailored for a specific device, then it takes into account width and screen of the device, and adding link would mess with the representation
Kai: support that Cache-Control: no-transform needs to be a strong statement
Rob: OK, but I still point that URI rewriting is a problem
francois: redirect doesn't solve the problem?
rob: probably, but not in all cases. There could be cases where various proxies in the path would proxy the HTTP 302 redirect back to the CT-proxy again. But this is only a technical problem to overcome, not a reason to not specify best-practise.
<DKA1> PROPOSED RESOLUTION: In the context of tasting content, if header comes back as cache-control no-transform, then CT proxies SHOULD change to transparent proxy operation (e.g. sending a http redirect).
Jo: I don't think we need to write this, because we already say that the proxy SHOULD be transparent "unless".
<jo> +1
Jo: I don't particularly object
to your resolution though.
... Agree we need to be clear.
RESOLUTION: In the context of tasting content, if header comes back as cache-control no-transform, then CT proxies SHOULD change to transparent proxy operation (e.g. sending a http redirect)
<jo> PROPOSED RESOLUTION: For no-transform responses and for https change text referring to "express user choice" to mean "case by case choice"
DKA: next proposed resolution to
resolve the issue within the next half hour?
... does that not imply that the user will be repeatedly
prompted with security warnings?
jo: the point is that blanket expression of user preferences should not be used here.
rob: insertion of a warning header by the CT-proxy in the page is enough?
francois: no, the CT-proxy must
not start transformation of HTTPS content without the agreement
of the user
... same thing with HTTPS links within HTTP pages
SeanP: Can the proxy keep the user's choice for the web site?
DKA: yes, it would be the middle
ground between blanket and case by case. The user could be
given different choices: only for this page, for the site,
...
... but that may be going too far in prescribing how the
CT-proxy may behave
Dom: I don't think we need to be explicit about that
francois: I note that what is in the guidelines for HTTPS rewriting is already the result of such a discussion on how to rephrase it so that we stay on this middle ground
DKA: OK, can we translate the HTTPS part to the Cache-Control: no-transform directive?
<jo> PROPOSED RESOLUTION: If the response includes a Cache-Control: no-transform directive then the response must remain unaltered other than to comply with transparent HTTP behavior and other than as noted below. If the proxy determines that the resource as currently represented is likely to cause serious mis-operation of the user agent then it may advise the user that this is the case and must...
<jo> ...provide the option for the user to continue with unaltered content.
<jo> PROPOSED RESOLUTION: If the response includes a Cache-Control: no-transform directive then the response must remain unaltered other than to comply with transparent HTTP behavior and other than as follows. If the proxy determines that the resource as currently represented is likely to cause serious mis-operation of the user agent then it may advise the user that this is the case and must...
<jo> ...provide the option for the user to continue with unaltered content.
DKA: Should we lower the
statement: "If the proxy knows better"
... ?
francois: I don't think so. The
response that is being altered may not be designed to be
displayed directly to the user. It could be the result of an
AJAX call for instance. I still support a strong Cache-Control:
no-transform directive.
... Adding an interstitial response would break existing
content
<DKA1> +1
DKA: I suggest that we move forward on this resolution, and that we use the Last Call period to get feedback from CT vendors and Content Providers. I suggest we use the Last Call to poll feedback.
<jo> a. how the CT proxy should react against specific user preferences when a site becomes more capable
<jo> b. that the default experience should be for the mobile experience
<jo> c. that blanket expression of preferences should be interpreted in the context that if an origin server can provide a choice of experience then it should do so
<jo> d. that CT proxies should, in restructured content, proivide links to alternative representations
<jo> e. how would a CP indicate to a CT Proxy that it offers user choice of representation?
<jo> f. allow users to change previosuly expressed preferences
francois: agree with the resolution, although I don't see the difference with what we already have in there.
RESOLUTION: If the response includes a Cache-Control: no-transform directive then the response must remain unaltered other than to comply with transparent HTTP behavior and other than as follows. If the proxy determines that the resource as currently represented is likely to cause serious mis-operation of the user agent then it may advise the user that this is the case and must .provide the option for the user to continue with unaltered content.
DKA: Another resolution we can take in the remaining 10mn?
Jo: [going through the above
list]
... On point e: we already resolved on the use of the Link
element with some reference somewhere in the document. Now we
need the following:
<jo> PROPOSED RESOLUTION: If the server has alternative representations then it should indicate this using link header/elements
<DKA1> +1
<rob> +1
Kai: What about POWDER?
RESOLUTION: If the server has alternative representations then it should indicate this using link header/elements
Jo: Then, how does a Content Provider advertise the fact that it proposes a choice to the end user between a desktop and a mobile representation?
dom: I think we should leave that as heuristics
jo: Overall, my feeling is that we're not going to make more progress on this subject today...
Kai: would appreciate if resolutions were not taken that quickly, especially when I have questions on the matter.
francois: summary of where we are
re. CT?
... I think we need to think carefully about it. It involves
complete rewriting of 3.2 in my view.
jo: I'd say the whole section
should disappear.
... Well have to wait for the new draft and see where the
direction goes.
DKA: On the agenda, I don't want
to move the points we missed this afternoon to tomorrow
morning
... I suggest that in order of priority that we move these
discussions into early afternoon on Wednesday.
Jo: I think that's fine.
DKA: Session adjourned