Re: [w3c/browser-payment-api] General Comment on Registry Design (#148)

On 2016-04-21 14:12, ianbjacobs wrote:
> @dret <https://github.com/dret>,
>     it seems that
>     https://www.w3.org/TR/2016/WD-payment-method-id-20160421/ is w3c's
>     first attempt to create a registry.
> It is true that we do not frequently do registries. But we have done
> some (e.g., xpointer registry).

true, i am sorry for that oversight. it's just a little different fomr 
IETF where the overall count of registries now is close to 2000. so they 
have an established way how to handle this.

> If the IG decides to keep some form of registry of payment method
> specifications or identifiers,
> then I agree we will need some sort of policy that explains how we
> manage the registry.
> For example, there may be some payment method specs that the WG
> publishes, and others
> that the community publishes and that we link to.

exactly. https://tools.ietf.org/html/draft-wilde-registries is where i 
am currently writing up design consideration that everybody thinking 
about defining a registry might want to ask themselves.

> But that gets to the point of why, to my understanding, we do not
> generally create registries: the Web is the registry.

the web may be used as a platform to host/manage the registry, but it's 
not a registry. questions such as design guidelines/review, 
documentation requirements/review, naming conventions/constraints, 
discoverability, and of how to manage non-URI namespaces are all not 
addressed by the web. the web offers countless ways how to implement a 
registry, but at first the important questions are what the registry 
should manage and how it should be managed, and then one can decide on 
implementation options.

> If payment method
> specs flourish, then I would expect that W3C would not try to track all
> that exist, but would simply maintain a list of those published by the
> WG (and that might be found on w3.org/TR for example).

that's one of the interesting questions of any namespace managed via 
registry: discoverability and registration requirements. such as the 
question of whether there's a two-tier namespace used such as the one 
established by RFC 5988, with short names managed by a registry, and 
non-registered extensions being represented by URIs, which optionally 
may make some documentation available at their address.


---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/w3c/browser-payment-api/issues/148#issuecomment-213164159