W3C home > Mailing lists > Public > public-bpwg-ct@w3.org > October 2007

[minutes] Minutes of Content Transformation Task Force Meeting 2007-10-09

From: Jo Rabin <jrabin@mtld.mobi>
Date: Tue, 9 Oct 2007 16:15:18 +0100
Message-ID: <C8FFD98530207F40BD8D2CAD608B50B473333A@mtldsvr01.DotMobi.local>
To: <public-bpwg-ct@w3.org>

Please find minutes of today's Task Force meeting at [1] and as text
below.

Jo

[1] http://www.w3.org/2007/10/09-bpwg-minutes.html


Mobile Web Best Practices Working Group Teleconference
9 Oct 2007

Agenda

See also: IRC log
Attendees

Present
    Rhys, Aaron, +46.3.17.47.aaaa, ROb, DKA, Andrew, shah, Magnus,
SeanPatterson, +035318aabb, Matt, jo
Regrets
Chair
    Rhys
Scribe
    rob

Contents

    * Topics
         1. administrivia
         2. Problem Statement
         3. Guidelines Document
    * Summary of Action Items

 

 

<trackbot-ng> Date: 09 October 2007

<Magnus> somebody tell zakim that +46.. is me

<scribe> scribe: rob

<scribe> scribenick: rob
administrivia

<j1> who is on list

Rhys: please sign up to CT mailing list [6] using link [8]

<j1> -> mailto:public-bpwg-ct-request@w3.org?subject=subscribe subscribe
to list
Problem Statement

<j1> [latest draft is 1f]

<Rhys>
http://www.w3.org/2005/MWI/BPWG/Group/TaskForces/CT/editors-drafts/Probl
emStatement/071008

Rhys: close to agreement on this
... Rotan had qn on performance overhead

Jo: belongs in statement of principles
... in the guidelines doc

Sean: where was the qn from Rotan posted?

Jo: wasn't posted to the list - that explains why you'd not seen it

Rhys: I'll respond to Rotan directly

<Rotan> thanks

Rhys: tracker shows open issue 221 Brian Sullivan

Jo: is relevant but had no response from Brian or anyone else about the
changes Jo's made

<j1> jo's comments on Bryan's comments

Jo: where it says suggested change from...to is response to this issue

Rhys: for things not in Problem Statement, do they naturally live in
Guidelines?

Jo: no
... Brian's change not justified. Brian hasn't responded to Jo's
response

Rhys: given time elapsed for Brian to respond inclined to propose
Problem Statement goes to wider working group on Thursday

<Magnus> +1

Rhys: but Brian doesn't seem to be on the CT mailing list

Jo: to become a Note, do we need to publish a Working Draft now anyway?
... most polite approach is report comments to Brian directly
... and hold-off working group distribution to a week Thursday

<jo> ACTION: Jo to forward his comments to Bryan to request feedback if
possible by Thursday [recorded in
http://www.w3.org/2007/10/09-bpwg-minutes.html#action01]

<trackbot-ng> Created ACTION-574 - Forward his comments to Bryan to
request feedback if possible by Thursday [on Jo Rabin - due 2007-10-16].

Rhys: 3 more open actins are complete

c/actins/actions/

<jo> PROPOSED RESOLUTION: Request BPWG to request transition of Problem
Statement to W3C Note - modulo no objections fromBryan

Rhys: given Nigel Choi's public feedback, should this be Working Draft
or a Note?

Jo: his comments taken into account

<Andrew> +1

<kemp> +1

+1

<jo> RESOLUTION: Request BPWG to request transition of Problem Statement
to W3C Note - modulo no objections from Bryan
Guidelines Document

Rhys: tracker http://www.w3.org/2005/MWI/BPWG/Group/track/products/12
... Nigel Choi's commented on how to taylor guidelines
... Jo raised qn of alternate representation discovery
... Jo also raised a shopping list of things to consider

<jo> [another thing to consider is -> http://tools.ietf.org/html/rfc3507
Internet Content Adaptation Protocol (ICAP)

<Andrew> +q

Rhys: so do we have a fairly complete list of material to begin
developing guidelines from?

<jo> raised by Novarra to Philipp at MoMo London last night]

Jo: firstly we need to encourage members to contribute more and more
quickly than so far
... second there are a wide range of use-cases and techniques to look at
and how existing servers will respond to them
... our feet are in mid-air if we can't kick the tyres on ideas,
experiment and know what will happen with proposed changes

Rhys: ie make sure we don't break things

Jo: worst thing would be for guidelines to recommend something that
breaks services. we have a lot of work and experimentation to do
... so discussion so far seems inadequate

andrew: I'm collecting Vodafone ideas together 1st before contributing
here

<jo> +1 to not causing any more confusion!

andrew: and agrees with Jo need to test and experiment these things

DKA: experimentation sounds a good approach

<Rotan> Question: would any proxy vendor be willing to work with origin
server vendor to do such experiements in live environments?

DKA: with similar controversial proposals before have put up demo sites
and had people test with lots of browsers
... byt in this case it's servers we have to test - and that's harder

c/but/but/

<jo> [think that dotMobi would be willing to help with this, think also
that there are some members that are well known for th scale of their
crawling!]

<kemp> i would be interested in helping with experiments where possible

<DKA> Is there an open source "reference platform" (such as Apache
running in a proxy mode) that could be used to trial any of the proposed
techniques?

Rhys: good starting point would be draft textual proposals that the
public can comment on
... which may result in narrowing down to a reduced number of proposals
to experiment with in practice
... have we pushed the boundaries yet or are there new ideas we haven't
seen yet?

Andrew: most of my ideas already voiced by others - and thanks for
everyone's contributions so far

Rhys: can we enumerate the ideas so far?

Jo: yes, but let's define the problem space 1st
... matrix of participants in the delivery chain and who's aware of
Guidelines and can adapt sensitively to them and who can't

Rhys: are we providing a number of independent guidelines that can make
things better and not break anything? or are there guidelines that only
work if more than one participant plays together?
... ie of the 8 possible scenarios are there 3 or 4 that are the
important ones?

<Andrew> also case of multiple CT proxies

Magnus: agree only a few important use-cases
... map out what each participant says to each other and the exceptions
(eg when a participant is completely unaware of that's going on)

<Zakim> DKA, you wanted to urge keeping the scope to when you have at
least one transforming element in the chain. and to wonder if there is
any difference between the case where you have

c/that's/what's/

DKA: urge against expanding scope

<DKA> cool

DKA: but do recognise the need for the few use-cases (eg multiple
transformation proxies)

Andrew: seen multiple transformation proxies already, essential to
address this

<SeanPatterson> +1 to not expanding the scope of the task force

Magnus: also essential is where origin-server can provide perfectly
suitable content for itself

Rhys: should concentrate on where participants are aware and can
interact appropriately

<Zakim> andrew, you wanted to discuss the case of ct proxies

<Rotan> Thought at first that "Vary: *" might do that (indicate support
for all context types) but also think that this would be valid for
blocking servers too (unfortunately). How can we indicate that server
gives *good* content for all devices? (i.e. not blocking)

Jo: this assumes dumb hosts and adapting proxies. And that non-aware,
non-adapting hosts are a significant use-case
... need to characterise behaviours of all the participants of varying
awareness to ensure we're not dammed if we do and dammed if we don't

Rhys: so we have to analyse all options with a view of if it still works
(or at least doesn't break anything) when participents in the delivery
chain are not aware of any of the Guidelines

Rob: good approach and the matrix Jo suggests is a good way to
concentrate the mind

Rhys: can someone enumerate the threads already proposed? in a mail to
the list?

Jo: I will

Rhys: thanks Jo

<jo> ACTION: Jo to enumerate techniques already identified with pointers
[recorded in http://www.w3.org/2007/10/09-bpwg-minutes.html#action02]

<trackbot-ng> Created ACTION-575 - Enumerate techniques already
identified with pointers [on Jo Rabin - due 2007-10-16].

<jo> s/somebody tell zakim that +46.. is me//
Summary of Action Items
[NEW] ACTION: Jo to enumerate techniques already identified with
pointers [recorded in
http://www.w3.org/2007/10/09-bpwg-minutes.html#action02]
[NEW] ACTION: Jo to forward his comments to Bryan to request feedback if
possible by Thursday [recorded in
http://www.w3.org/2007/10/09-bpwg-minutes.html#action01]
 
[End of minutes]
Received on Tuesday, 9 October 2007 15:16:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:10:36 GMT