An update

[Andy Buckingham]

Hi Team,

Yesterday we had another productive meeting with Chris Lowis and Sean
O'Halpin from the BBC and myself from Global Radio.

Chris and Sean presented the latest revisions to their previous
RadioTAG interaction model proposal. This work is coming along nicely
and the previously discussed auth process is now well proven through
their models and testing in several different states:

1. No tag service, not supported (SRV record lookup fails)
2. "Anonymous" tagging - nothing is ever stored on the server allowing
very simple implementations
3. "Out of box" experience - client obtains token from service
provider that allows tag submission and storage without the user
logging in or register, provider will remember (a limited number of)
tags, tags can only be viewed via that client
4. "Paired" - client is paired with user account, stored indefinitely,
view tags on client/web/other client

To ensure a fairly brief e-mail, I'll let Chris and Sean to reply to
this thread in the coming days with revised diagrams which explain the
current state of the proposal.

The only major point of change from previous revisions was that any
tag submission response body will now have a single item Atom XML doc.
This allows us to include an image and a bit more details so that
"Anonymous" tagging can offer the same experience as "Out of the box"
and "Paired". As an "Anonymous" device never gets a token, it will be
forbidden from using the normal service proposed to list all tags and
no tags are stored anyway, so the only chance to deliver meta data to
the client is via the first response. All other devices will likely
show this as a notification, but utilise the updated events feed
service for populating any kind of "view tags" menu. Another benefit
is that any notification for the delivery of a tag can now also
include a link, image and title allowing rich display devices to
provide an improved experience.

We also revisited the close proximity of our auth method with that of
the OAuth2 project and whether, should this proposal seek approval as
a finalised specification for RadioTAG, we eventually put forward the
auth section of our work for discussion with the OAuth team as a
proposal for limited device authentication (which was previously
removed from OAuth due to them finding a lack of suitable use case!)

More importantly, there was an agreement that we need to tidy up our
vocabulary of terms to describe values, states and methods in the
model. There is a plan to review OAuths terminology and then, via this
group, try and iron out a better naming convention for ease of
understanding.

Finally we have discussed updated timelines for the project and the
BBC are due to deliver their finalised proposal within the month. I
plan to review and subsequently produce a draft specification document
for moderation by this team shortly after that, with a target of late
August to finalise and release the RadioTAG 1.0 specification.

If you have an interest in the development of the final 1.0
specification and have been casting a casual eye over recent messages,
I would urge you to begin reviewing messages over the coming weeks as
closely as possible as we rely on feedback from the group to spot
errors we've missed or identify any problems for those planning to
implement the specification. If we do not receive reasonable requests
to review or amend areas of the proposal in a timely fashion we will
move on towards drafts and ultimately publication. This is your chance
to get involved and influence the specification.

Thanks for your time and I look forward to hearing from you.


Andy.