13:56:20 RRSAgent has joined #apps 13:56:20 logging to http://www.w3.org/2016/11/23-apps-irc 13:56:57 Ian has changed the topic to: 23 Nov agenda https://lists.w3.org/Archives/Public/public-payments-wg/2016Nov/0043.html 13:57:00 agenda: https://lists.w3.org/Archives/Public/public-payments-wg/2016Nov/0043.html 13:57:04 Chair: Ian 13:57:33 present+ 13:57:36 present+ Tommy 13:57:39 present+ Michelle 13:57:42 present+ Jenan 14:01:11 rouslan has joined #apps 14:01:16 present+ Rouslan 14:01:57 present+ Frank 14:02:06 stan has joined #apps 14:02:47 mbu has joined #apps 14:03:26 agenda -> https://lists.w3.org/Archives/Public/public-payments-wg/2016Nov/0043.html 14:03:29 present+ Stan 14:03:40 topic: Next meeting 14:03:42 30 November 14:03:55 Update on ACTION Ian 20161123: Look into a new meeting time based on who is participating in the call. Status: Not done yet. 14:04:29 topic: Implementer updates 14:04:42 IJ: Any updates since last week? 14:04:54 Rouslan: Samsung folks have set/get manifest 14:05:00 ..you can run those functions in service workers 14:05:08 ...in Chrome Dev 14:05:31 ...these are two functions in the payment app spec to register a payment app 14:05:41 ...there is a label and icon for the app, and a list of payment options 14:05:48 ..it's behind a flag 14:06:08 frank has joined #apps 14:06:16 chrome://flags/#enable-experimental-web-platform-features 14:06:30 https://github.com/w3c/webpayments-payment-apps-api/issues/69 14:06:40 What does "icon" mean? 14:06:57 IJ: any insights since last week? 14:07:10 Rouslan: Web App manifest uses URLs for icons...I guess that's what we are going to do 14:07:24 https://www.w3.org/TR/appmanifest/#icons-member 14:07:40 ...Samsung is really pushing us forward on web-based payment apps 14:08:27 whenever possible, let's use the same primitives as elsewhere in specs, like webapp manifests. 14:08:33 http://caniuse.com/#search=manifest 14:08:48 rouslan: I have read the code for web app manifest, which is fairly stable 14:08:56 ...I've seen apps use it to change the color of the address bar 14:09:12 (Seems Chrome implements it, but not widely implemented yet) 14:09:19 regrets+ adamR 14:09:20 ..it's also used for adding web sites to your home screen 14:09:45 ...I think we should do what they do for icons (list) 14:11:10 +1 copy def 14:11:28 Proposed: Copy web app manifest definition for "icons" (and update the spec to make it a list) 14:11:35 +1 14:11:37 +1 to copy 14:11:43 +1 14:12:00 so RESOLVED. 14:12:19 ACTION: Update the payment app spec to copy web app manifest term for icon_s 14:12:44 IJ: Any other observations from get/set work? 14:12:47 Rouslan: Not yet 14:13:13 Topic: Any Learnings from Web Share API? 14:17:50 IJ: See agenda for summary; next steps IMO will be UX challenges for Web-based payment apps 14:18:06 Topic: Issue 37 14:18:11 Can payment apps cancel payment flow? 14:18:17 https://github.com/w3c/webpayments-payment-apps-api/issues/37 14:18:34 https://github.com/w3c/webpayments/issues/128 14:18:52 IJ: Any new insights about cancelation? 14:18:56 q+ 14:19:00 tommy: Adam's comment seems reasonable 14:19:17 Tommy: I think I raised this pre service worker 14:19:18 ack rous 14:19:31 rouslan: Can we let the user select a different app if one of their apps fails? 14:19:40 q+ 14:20:13 ack rouslan 14:20:30 rouslan: there are two points of failure detection: 14:20:53 1) browser hands payment request to app; app comes back with failure....browser knows that payment app doesn't work and the browser could should the user other matching payment apps 14:21:05 ...this is how it works on Android today when AndroidPay today 14:21:23 ....or, e.g., if the user cancels CVC, the user can still pay with another credit card 14:21:55 2) the other failure mode is the payment app succeeds, hands JSON to merchant who tries to use the data and then FAILS. 14:22:18 ...in this case, the merchant calls instrumentResponse.complete('fail') at which point the browser displays an error message that the transaction failed 14:22:46 ...in this case the only user interaction is to click "Close" on the error message and then the merchant can create a new payment request 14:22:48 ...or fall back 14:22:53 q? 14:23:21 https://w3c.github.io/webpayments-payment-apps-api/#sec-payment-app-response 14:23:43 +1 variety 14:24:08 q+ 14:24:20 ack rouslan 14:24:31 IJ: Seems like at a minimum that the spec needs to not prohibit "retry" 14:24:50 ...whether we encourage it (or even require) is another question (and premature to answer) 14:24:53 +1 to not prohibit "retry" 14:24:54 rouslan: +1 to variety 14:25:06 "The user agent receives a failure response from the payment app through rejection of the Promise. The user agent MUST use the rejection reason to reject the Promise that was created by PaymentRequest.show(). " 14:26:27 IJ: So let's create an issue about user agent behavior when payment app fails. 14:26:59 ACTION: Ian to add a pointer from the spec to this issue 14:27:31 I have made the request to generate http://www.w3.org/2016/11/23-apps-minutes.html Ian 14:28:16 Tommy: Now that we've discussed this a bit, I think what would be useful would be for an app to have a graceful way of saying "I give up" 14:28:39 ...e.g., what if the app wants to cancel in a pretty way, could we make it nicer than sending back "invalid data" 14:29:32 ..maybe we should at least recommend some way to cancel 14:30:19 ACTION: Tommy to write a proposal for section 10.4 around issue 37 to consider (1) graceful failure in the case of cancelation and (2) mention of retry 14:30:27 I have made the request to generate http://www.w3.org/2016/11/23-apps-minutes.html Ian 14:30:39 Topic: Other pull requests 14:30:53 - WebIDL fix pending 14:31:06 - AHB commented on my edits and I'm trying to work through them with him 14:31:59 Topic: Wrap-up 14:32:07 IJ: Any other topics? 14:32:43 IJ: Frank, need any more information to play with payment apps? 14:32:53 Frank: I will reach out to Rouslan separately. 14:33:57 Have a good week! 14:34:00 rrsagent, make minutes 14:34:00 I have made the request to generate http://www.w3.org/2016/11/23-apps-minutes.html Ian 14:34:06 rrsagent, set logs public