See also: IRC log
trackbot, start meeting
<trackbot> Date: 01 August 2012
<scribe> Scribe: Josh_Soref
<scribe> Agenda: http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0056.html
Jo: present, Jo, darobin, Josh_Soref, andrew_betts
... jfmoy
... bryan
... tobie
kai: Kai Fritz, vodafone
Jo: who else?
... is the w3c contact?
darobin: probably not
[ Chairs negotiate ]
<Jo> http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0055.html
Jo: the summary took 2 days to produce, which is rather long for me
... i'd like people to make sure the resolutions were recorded correctly
... particularly the after lunch resolutions on day 2
<Jo> PROPOSED RESOLUTION: Adopt the draft minutes from this meeting with the amendment that the meeting took place on 25 and 26 June not 28 and 29 June as stated
darobin: sounds good
tobie: ian jacobs asked for an executive summary
... it would be useful for w3, for us internally, and maybe for the w3c blog
Jo: sounds reasonable
... since it is rather lengthy
RESOLUTION: Adopt the draft minutes from this meeting with the amendment that the meeting took place on 25 and 26 June not 28 and 29 June as stated
<Jo> ACTION: Rabin to summarise the summary and post on CG blof [recorded in http://www.w3.org/2012/08/01-coremob-minutes.html#action01]
<trackbot> Created ACTION-46 - Summarise the summary and post on CG blof [on Jo Rabin - due 2012-08-08].
<Jo> > http://coremob.github.com/level-1/index.html Current Draft Level 1
tobie: i'm going through the updates on the spec
... i've gone through most if not all of the actions specific to it
... there's a couple i have not done yet
... mostly because i have further questions about them
... or a bit of work to do
... but mostly i've closed the related actions
Jo: we'll step through your actions
... i've moved your actions to pending-review
... anything you'd like to raise?
tobie: not really
... i need to do some research before bringing them back to the group
... i don't want to bother the group before
Jo: anything else on level 1 current draft?
[ no ]
Jo: thanks tobie
<Jo> http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0031.html
Jo: AT&T offered
... to redraft level 0
<darobin> http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0038.html
Jo: my action-2 was to propose something different about level 0
... a way that would allow it to proceed in some orderly form
... we baselined features based on Matt Kelly's list
... there's also a discussion based on Mobile Web Best Practices
<Jo> http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0043.html
Jo: here's my action-2 + action-4
... i've joined those two things together
... i'm proposing reconsidering what level 0 is
... to focus on a small set of UCs from Matt Kelly's definitions
... i wonder if anyone has any views?
bryan: i sent to the list our intent here
... we're ok with creating a document based on a wide range of devices/browsers
... for a defined set of features
... or taking a UC based approach
... analyze the features
... i think the UC based approach is interesting
... but i think people need information about what's available in browsers
Jo: isn't that exactly what derailed progress on Level 0 initially?
... i'd rather not capture the initial state
... but capture definitional
... the group has reached a point where we can say Level 0 is aspirational
Josh_Soref: I think that we've reached a point where calling Level Zero is going to be a third rail
... so if someone wants to draft something they shouldn't mention level or zero
... there are sites that list features and support on browsers
... and don't do compliance, but some sites do things like it
... while we don't have test suites that do compliance, we should defer to other sites (e.g. caniuse.com)
<bryan> public sites e.g. caniuse provide support info inconsistently and incompletely
Josh_Soref: it doesn't seem like a very good use of resources for us here to do that
... and I'm okay with something aspirational or definitional, but it should be somewhat narrow
... working from Matt Kelly's list isn't unreasonable
... but it should be limited in scope
<bryan> I do support the strategic focus either on verification or aspiration
Josh_Soref: and use meaningful names in its title
tobie: one of the mistakes we made when we announced level things
... was not to consider the audience of these documents
... we got feedback from the implementers
... who are the audience
<bryan> but we need to be clear which we are trying to achieve
<chaals> [seems that it would make more sense to spend effort providing information so caniuse.com or whatever you like most is more accurate and complete. It's the same work that would be required anyway]
tobie: was that they are *REALLY* not interested in documenting things we're moving away from
... but would rather document things we're moving towards
... this is something to keep in mind when talking about level 0
... if the main target is developers
<bryan> Tobie, what does "moving away from" mean?
tobie: then produce a document like that which would make implementers less cringy
bryan: what did you mean by "moving away from"
tobie: a document containing "EcmaScript 3"
... where implementers are moving to EcmaScript 5
... implementers are moving from HTML4 to HTML5
... that's the crux around documenting anything around the main market of mobile devices
bryan: i think i understand what you said
... i'm not sure i get it
... as i dropped on irc
... i think the group needs to decide if it's going to document what's supported
... or to establish aspirations
... if you're looking at aspirations, you can look at what's around the corner
... i don't think caniuse.com/browser scope
... is complete or consistent
... the methodology is extremely fragmented
... it's extremely difficult to use as a developer
... one of our main focuses is to draw that together into a place that's automated, easy to consume
Jo: i don't think it's this group's job to repeat what's done by browserscope/caniuse
... i'd like us to agree to move on if that's a consensus view
... it seems the group has proven by a non-existence proof
... that the baseline doesn't exist
... and to move to an aspirational thing
... defining "if it has this, this and this" then it is a mobile application [host?]
tobie: if we're interested in giving a good picture of the world today
... this could be done by publishing the results of how well existing browsers fair on the level 1 spec
<bryan> agree, we are not trying to replace caniuse etc, but trying to promote consistency or agreement on what should be consistent
<Jo> PROPOSED RESOLUTION: This group does not intend to reproduce the results of Browserscope or CanIUse
tobie: if we have a good test suite, then the current state
... of browsers is whatever percentile you're interested in
... say 90% of the browsers get as results
<bryan> how do developers fill the gaps then?
<Jo> PROPOSED RESOLUTION 2: There is no test suite below Level 1
<bryan> above the 85-90%
jfmoy: i agree with bryan
... on these concerns
... i'm surprised with discussions on caniuse / similar tools
... the results they provide is far from being
... i don't have a good level of trust in them
... we found discrepancies when comparing with our own internal testing
<Jo> PROPOSED RESOLUTION 3: Level 0 (if it exists at all) exists as a definition only what it means to be a mobile web app
darobin: as a data point
<darobin> http://w3c-test.org/framework/app/report/matrix-maker
darobin: matrix maker is being built
... so you could pick browsers, specifications
... to give you a matrix of support for various browsers/specifications
... and you'll be able to save that report
... it ought to be available by the end of the month
<bryan> that will be useful, and CoreMob could then based upon that data issue some statement as to the baseline
<Jo> PROPOSED RESOLUTION: This group does not intend to reproduce the results of Browserscope or CanIUse
<bryan> +1
<darobin> +1
<jfmoy> +1
dan: i'm blocked from irc
Jo: have you tried http://irc.w3.org/ ?
<Jo> PROPOSED RESOLUTION 2: There is no test suite below Level 1
RESOLUTION: This group does not intend to reproduce the results of Browserscope or CanIUse
<bryan> but based upon some testing somewhere, will the group establish what t thinks *is* the current state of the "Core Mobile Web" support?
Jo: bryan, no, i don't think it will do that
... it might define what it means to be a mobile web application
bryan: and explain why/when it matters
Jo: i think defining the difference between a mobile web site and a mobile web application
bryan: defining the state/aspiration
... the output needs to be something that is useful for developers
... if things aren't supported, but are filled through libraries
... that has a place in the landscape
Jo: i think that could be discussed on list
bryan: i think that should be noted
<Jo> PROPOSED RESOLUTION 2: This group will not produce a test suite that is below the aspirational Level 1
<bryan> how polyfills play a role in current levels of support across browsers, and that they are a fact of the "mobile web" experience for developers
dan: for level 0, if we only have documentation
... are we recommending developers use browserscope/caniuse?
Jo: we're saying we want to complete a range of features across a range of browsers
... the point of the group is to look forwards
... maybe slightly in the future being december
... any test suite that would cover level 1 will cover the features that are important today
<bryan> I think one can only look forward and make sense of what is seen, by knowing the current point from which the observation is made, e.g. is the camera ahead or behind us?
Jo: aside from a discussion of what is in a level 0 test suite
dan: will we include all the specific features from level 0 into level 1?
Jo: i'm not sure how one would do a test suite of HTTP
tobie: i've already folded all the non-controversial parts of level 0 into level 1
... i did that before the f2f
<bryan> Folding level 0 into level 1 is I think a good compromise approach, if we want only to look forwrd
dan: what do you mean by the controversial stuff
<bryan> appcache etc?
tobie: there were a number of features included in the early draft of level 0
... some accidentally
... which made a lot of people very unhappy
... these included things that were encumbered by patents
... and vendor prefixed css features
dan: for prefixed features
... i thought we agreed to test nonprefixed
Jo: we're going off topic
... we agreed at the F2F to only focus on spec'd features
<Jo> PROPOSED RESOLUTION 2: This group will not produce a test suite that is below the aspirational Level 1
Jo: i'd like to return to the conversation
<tobie> +1
<darobin> +1
Josh_Soref: +1
RESOLUTION: This group will not produce a test suite that is below the aspirational Level 1
<bryan> +1 given that level 1 is as described by Tobie
<Jo> PROPOSED RESOLUTION 3: Level 0 (if it exists at all) exists as a definition only what it means to be a mobile web app
<tobie> bryan: http://coremob.github.com/level-1/index.html
Jo: we define a mobile application
<bryan> thanks, I will review the current level 1 spec
Jo: as something which has a degree of autonomy
... and doesn't require online access for extended periods in order to do whatever it does
... whatever came out of it would be a minimal spec
... http capable, bootstrap from the web
... is it meaningful to call that level 0?
<bryan> I think it's a useful definition
dan: i think that's too minimal
Jo: what i'm proposing is that we review the list of features that matt kelly came up with
... if it can do this, you can't call it a mobile web app
tobie: there's been a notion of dropping leveling
<chaals> [Why not let Bryan have a go and see if he writes something that convinces us? It might be quicker for everyone else) than deciding a priori whether there is something people might accept and value]
tobie: since we don't have a level 0, and don't have a notion of a level 2
... let's release level 1 as "Core Mob ..." or "Mobile ..."
... and then see if we want to do things as yearly snapshots
... or levels
<bryan> I would like to see some results from whatever testing is done, either in the Testing IG or elsewhere, factored into the discussion on defining "what a mobile web app is" as what is supported clearly limits what an app "can be"
tobie: maybe later have a modern web app definition
... have something distinct from level 1
Josh_Soref: +1
dan: i agree
<Jo> PROPOSED RESOLUTION:Rename proposed definition doc as "What is a Web App" drop the whole idea of Level 0 and rename Level 1 as ScoreMob 2012
<Jo> PROPOSED RESOLUTION:Rename proposed definition doc as "What is a Web App" drop the whole idea of Level 0 and rename Level 1 as CoreMob 2012
<darobin> +1
<dan> +1
Jo: any objections to dropping level 0?
<bryan> +1 (sorry, don't want to belabor a difficult discussion, the group should move on)
Josh_Soref: +1 from me for dropping it
RESOLUTION: Rename proposed definition doc as "What is a Web App" drop the whole idea of Level 0 and rename Level 1 as CoreMob 2012
Jo: is there support for doing a small definitional document on what is a web app?
... bryan, would you volunteer to do it?
... you volunteered for something different
<bryan> we will have opportunity to do what is needed, once we have the data to do so - any document prior to that would be preliminary
<andrew_betts> +1 to the need for a clearer understanding of what a 'web app' is.
<Jo> ACTION: Sullivan to draft a chapter outline of "What is a Web App?" [recorded in http://www.w3.org/2012/08/01-coremob-minutes.html#action02]
<trackbot> Created ACTION-47 - Draft a chapter outline of "What is a Web App?" [on Bryan Sullivan - due 2012-08-08].
Jo: we won't have time to do action bashing
Jo: three things to discuss
... 1) shall we have another call in 2 weeks' time?
jfmoy: I would prefer a month
+1
<Jo> PROPOSED RESOLUTION: Further meeting in 1 month
RESOLUTION: Further meeting in 1 month
<Jo> ACTION: Jo to arrange September call [recorded in http://www.w3.org/2012/08/01-coremob-minutes.html#action03]
<trackbot> Created ACTION-48 - Arrange September call [on Jo Rabin - due 2012-08-08].
<bryan> jo, please send me what you described as the basic characteristics of a "mobile web app", and we will flesh it out
Jo: 2) noting the dates of the f2f are Oct 2-3, with logistics to follow, any further questions? I'll circulate details
... AOB?
... then let us thanks the FT
<Jo> PROPOSED RESOLUTION: Coremob thanks the FT for hosting the voice channel for this call
<dan> +1
RESOLUTION: Coremob thanks the FT for hosting the voice channel for this call
[Adjourned]
<Jo> [Following the adjournment co-Chair realised that he'd been dreadfully rude in not thanking the Scribe]