See also: IRC log
<scribe> Scribe: Ian
https://www.w3.org/blog/wpwg/2017/05/18/payment-handler-api-first-public-working-draft-published/
IJ: Upcoming implementation plans or ideas?
rouslan__: In Chrome...we are
developing the API for Android and Desktop. Our desktop PR API
implementation has not yet shipped, so we are focused on
Android for now.
... we are looking into how to properly open the window
... so that it looks good for the payment app that is a website
(on android)
... we have some ideas for that
... on desktop, we haven't shipped PR API yet but are close to
doing so; showed at Google I/O
... we are having UX discussions about how to best present
payment apps on desktop. What was mentioned in Chicago still
stands I think - the payment app web page will be shown in a
dialog window where the user would do the payment
IJ: Any word on desktop implementation availability?
rouslan: Look for it in stable release in Aug/Sep
IJ: Would be great to hear other plans
frank: We are happy to start experimentation as soon as there is support in the browser for web-based payment apps, which sounds like it is coming soon.
IJ: Any critical path features that we need to deal with to do experiments?
rouslan__: We need to automate
web platform tests that involve user interactions
... Marcos has a proposal that involves a payment method
specific to testing environment.
... that way we can test conformance of browsers.
... there seems to be consensus on this approach for PR API;
perhaps we can do something similar for Payment Handler
... e.g., reserved payment method for a testing environment
that could be used for automation testing.
IJ: Anything that could be usefully done to prepare for implementation testing?
IJ: I am in touch with people
suggested by Marcos
... Ali Alabbas, Joshua Bell
... meet with them later today
... perhaps need an asyncmaplike
https://heycam.github.io/webidl/#idl-maplike
https://github.com/w3c/payment-handler/issues
IJ: Any ones people want to look at in particular?
https://github.com/w3c/payment-handler/issues/157
rouslan__: I think that what implementers want is to keep PR API shape as-is and use the algo text to define how exactly filtering happens.
IJ: It's one thing to say "the algo says this" and another to say "where does it really happen -UA or payment app?"
rouslan: I think jury is still
out.
... weigh privacy v. capabilities of payment handlers
... I think that payment handler API will be more specific than
just "filtering happens here".
... but ack rous
IJ: Are we prepared for a canMakePayment() type approach for payment apps (which we said "no" to previously)?
rouslan__: I think we are going
to compare details...don't think we will call any
functions
... having said that, we are still in the early stages of
experimentation with Payment Handler API and payment app
writers
IJ: Should people on this call talk to you about those experiments?
rouslan__: Please do
IJ: Please contact Rouslan to help address this issue through implementation experience.
https://github.com/w3c/payment-handler/issues/128
(Maybe those are issues 2 and 3 now)
IJ: Any updates from implementers since the FTF?
rouslan__: We've not heard from
other app providers besides Klarna on this use case.
... but leave open
https://github.com/w3c/payment-handler/issues/74
https://github.com/w3c/payment-handler/issues/74#issuecomment-291548795
IJ: What are plans for browser-assisted support for recommendations of mobile payment apps?
rouslan: Have had some
discussions but not implemented...
... we think that recommended payment apps could work better on
web than native mobile
... loading a payment app would be useful to a user
... so we think that recommend payment apps might be useful ...
the spec may not need to change to support it
... but we haven't really designed or implemented of this; just
some ideas for now
IJ: Has this come up?
rouslan__: We keep hearing this
might be a requirement. We might be adding this to the native
android payment app spec
... not sure yet, though
... I would not be opposed to adding an abort() event to
payment handler...some nativepayment app developers have asked
for this
alyver: Beyond a certain point, abort may not always be successful. is the expectation that payment apps be able to handle abort() or do a refund for example?
rouslan__: I think the
expectation is that the event is fired into the payment app if
the merchant site requests an abort (e.g., timeout or no more
tickets available)
... so payment app could do different things depending on the
user's activity...e.g., if user is just starting to type
info
... but if payment app is processing or has completed
transaction...hten the payment app can signal that the abort
failed.
... this gives the merchant site some idea of whether their
abort request went through.
IJ: +1 to basically saying this
is a signal, rather than having conformance requirements
specifically for payment apps
... Should we wait for rouslan mobile experience before
spec'ing out?
<alyver> +1 to it being useful
https://github.com/w3c/payment-handler/issues/117#issuecomment-291522464
IJ: Rouslan, anything we can do to help you out?
Rouslan: Await further updates!
<rouslan__> +1
<alyver> +1
Proposed two weeks - 6 June
Next meeting: 6 June