Re: Concerns about Core Messages specification

On 07/21/2016 07:51 AM, Adrian Hope-Bailie wrote:
> I would prefer we define the HTTP API messages in the HTTP API spec 
> and then abstract these and the browser API messages into a separate
>  document when it is clear that is a good idea.

At present, the PaymentRequest and PaymentResponse messages (the data
model) are effectively the same between the Browser API and the HTTP API
(they carry the same basic sort of information). The data models are
already 80%+ aligned because we're actively trying to do that.

This is a FPWD, we're not saying the spec is a "good idea" yet. If that
was the bar for FPWD, our company would have objected to a number of
other specifications this group has published. An FPWD is saying "here's
what some of us are thinking right now and we're looking for feedback".
We can put issues, such as the one you mention, in an issue marker.

> Why would we do that if the only implementation of the core messages 
> is the HTTP API?

Because it's good spec design to separate messages from protocol
implementation even if you only have one protocol. This enables messages
to be re-used (in part or in whole) in other protocols. Case in point:
schema.org

schema.org was put together for information indexing (e.g. knowledge
graph, open graph) purposes by Google, Microsoft, Yandex, and Yahoo. The
schema.org messages are now used across a variety of different protocols
(SEO, email, credential publishing, geospatial databases, healthcare
databases/protocols, etc.).

This happened because they decoupled the spec for the messages from the
protocol.

> I believe it is premature to publish the Core Messages spec as it 
> sends the wrong signal about the current consensus of the group. I 
> don't believe the group is in consensus that there should a shared 
> data model for all APIs. Publishing Core Messages suggest that there 
> is.

It doesn't if we add a note that states that clearly.

Here's some suggested text:

ISSUE: Is a shared payment messages data model useful to the ecosystem?

The Web Payments WG is currently debating whether a shared payment
messages data model is useful to the Web Payments ecosystem. At present,
the HTTP API uses the data model in this specification. The Browser API
messages significantly overlap with the message definitions in this
specification. Should the group strive to share message definitions
between both protocols, should the messages be specific to the protocol,
or use a hybrid approach where some message components are optional?

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny, G+: +Manu Sporny)
Founder/CEO - Digital Bazaar, Inc.
blog: The Web Browser API Incubation Anti-Pattern
http://manu.sporny.org/2016/browser-api-incubation-antipattern/

Received on Thursday, 21 July 2016 14:34:13 UTC