Device APIs Working Group Teleconference

07 Sep 2011


See also: IRC log


Robin_Berjon, Frederick_Hirsch, Josh_Soref, Anssi_Kostiainen, Dominique_Hazael-Massieux, Kihong_Kwon, SungOk_You, Cathy_Chan, Bryan_Sullivan, Niklas_Widell
Robin_Berjon, Frederick_Hirsch
Josh_Soref, darobin


<trackbot> Date: 07 September 2011

<Josh_Soref> Scribe: Josh_Soref


<fjh> Webinos spec publicly available, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0012.html (Claes)

<fjh> request to TAG re Javascript API - HTTP/REST, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0014.html (Frederick)

<fjh> proposal with request for comment, on referencing HTML5, http://lists.w3.org/Archives/Member/member-device-apis/2011Sep/0000.html

Minutes Approval

<fjh> http://lists.w3.org/Archives/Public/public-device-apis/2011Aug/att-0105/minutes-2011-08-31.html

RESOLUTION Minutes for last week approved


<fjh> editorial clarifications, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0016.html

<fjh> Alternative events proposal, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0005.html

<fjh> event handling, no event when listener registered, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0035.html (Anne)

<fjh> synthetic events, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0028.html (Anne/Robin)

<fjh> most recent, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0044.html

<AnssiK> http://dev.w3.org/2009/dap/system-info/battery-status.html

AnssiK: I've tried to address all of the issues from the list in the latest ED
... it addresses the side effect issue by side stepping it
... it initializes the battery details as part of the constructor
... using tasks
... it uses the new event constructor and references DOMCore
... events are now cancelable and bubble
... clarified conditions of battery low, and critical based on Kihong_Kwon's feedback
... I dropped start/stop methods from the proposal
... start is handled by constructing the event
... stop can be optimized by UAs,... when there are no listeners, they don't need to fire it
... but that's an implementation detail
... the latest ED is there
... I reworked all the prose in the examples to reflect the changes
... it's a very small change from the developer's perspective
... instead of exposing the global window object, the info hangs off the battery object
... this feels cleaner
... Please review. I'd like to get implementation feedback
... I was in a hurry, so there are probably bugs/typos in ED

<fjh> need to fix missing DOMCORE reference

<Zakim> dom, you wanted to discuss constructor vs well-known instance

AnssiK: so please review it and send feedback

dom: AnssiK , first, thanks for all the work you've put in
... i like the new approach very much

<darobin> +1 to Anssi doing great work, should become editor of other specs :)

dom: i have a question about it, that i just sent to the list
... do we really want to allow creating as many Battery Event sources as a developer wants?

<fjh> dom's question on the list -> http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0047.html

dom: instead of making it a single well known thing?

AnssiK: the main problem is deciding when to enable the battery source

dom: i'm not sure what you mean by emitting

<darobin> [in fact we could require it to be a singleton]

AnssiK: the problem with hanging the data off the global object

<fjh> section 3.2, http://dev.w3.org/2009/dap/system-info/battery-status.html#constructor-1

AnssiK: is that there's no way for an implementation to optimize away gathering data

<dom> window.navigator.battery.start()

dom: i'm confused about why a developer would need more than one instance

AnssiK: the dictionary should show up in section 4
... in some browsers, it doesn't do the right thing, so i need to monkey patch things
... is that right darobin ?

<AnssiK> dictionary BatteryStatusEventInit should show up in Section 4

<AnssiK> I've monkeypatched it in and it's not working in all browsers it seems

<fjh> anssi notes perhaps WebIDL dictionary support needs to be in ReSpec

<AnssiK> so if it does not show up, you should reload the page

dom: there's an indirect request to add [Web IDL] Dictionary support to ReSpec

<Zakim> fjh, you wanted to note possible confusion on mail list

fjh: has the current discussion / ED resolved the initialization confusion on the list?
... involving side effects

<dom> (should we close ISSUE-113 or is that premature?)

fjh: Bjoern Hoehrmann

<Zakim> Josh_Soref, you wanted to note that implementations can cheat and make it a singleton and to talk about how developers use modules which don't talk to eachother

<scribe> Scribe: darobin

Josh_Soref: there is no requirement that new Foo() to return a different object
... we could return a singleton
... this is easy to implement in JS, plenty of implementation strategies, little cost.
... If you have three modules from different groups, they may want to do different things
... they'll each instantiate an object
... if they do that, each module is not going to coordinate over getting that object
... you could get that from navigator, but there are reasons not to

<Zakim> darobin, you wanted to insert an agenda item about workshop

[note that if a module 1 calls stop it would stop for all other modules -- assuming there is stop()]

<scribe> Scribe: Josh_Soref

darobin: do we want to publish a heartbeat for Battery?

Josh_Soref: +1 for a snapshot

<dom> [let's wait to get some further feedback on the new approach?]

Josh_Soref: i'll try to proofread it today or tomorrow

<fjh> +1 to week for Anssi to edit and for some review

Future of Offline Applications

darobin: there's a workshop coming up
... however so long as the workshop has not been officially announced
... everything is subject to change
... the topic potentially covers Widgets/AppCache

<bryan> This is what I proposed for a TPAC discussion could we have it that week?

darobin: the tentative date is the Saturday after TPAC
... tentatively at the Vodafone offices near SC.CA.US

bryan: why can't this be @TPAC?

darobin: i dunno

<fjh> ... e.g. on Wednesday

darobin: i think the idea is that the workshop expects to use an entire day
... if you have specific questions, i think the best person to ask is dan_a

[ darobin looks for location ]

bryan: i thought that Widgets were rather stable

<darobin> Redwood City, CA

darobin: you can also take issues to matt warner

dom: given the timeline, i think it's too late to change the details of the logistics
... given the requirement of a 8 weeks advance notice

<Zakim> Josh_Soref, you wanted to ask about a saturday moratorium

[ action dom to send feedback about scheduling on weekends ]

Feature Permissions

<fjh> revised example, http://lists.w3.org/Archives/Public/public-device-apis/2011Sep/0006.html (Wonsuk)

fjh: there was a little work revising the example
... it's unfortunate that lgombos isn't here


fjh: is there anything to cover here?

[ Adjourned ]

RRSAgent: bye

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2009-03-02 03:52:20 $