API Refactoring and TAG minutes

Please find the minutes for our API refactoring discussion [1] on
Tuesday an initial TAG review of our API [2] and an Auto+TAG session [3]
on Wednesday.

The impression I had during the refactoring call is: 

* We clearly ruled out a low level Sensor API approach
* The event versus subscribe model was not clearly settled but clear
need for the data broker to be able to detect when signals data are no
longer needed to avoid CPU and memory consumption issues
* There is strong interest in a Web Sockets RESTful over the current
WebIDL based API.  There were some who felt this might be a setback and
affect our timeline, others believe we could make the transition rather
quickly.  During the Auto+TAG session on Wednesday [3] there was more
momentum behind sockets.

Philippe,  I am bringing this to your attention since it would be best
in my opinion to be consistent with the LBS API and should we conclude
to redesign around Web Sockets, you won't have to delve further into
WebIDL.

The initial TAG review [2] has some excellent detailed feedback but
obviously we should settle on WebIDL vs Web Sockets topic before
addressing the points they raise as some may no longer be relevant.

[1] https://www.w3.org/2016/02/09-auto-minutes.html
[2] https://pad.w3ctag.org/p/03-02-2016-minutes.md
[3] https://pad.w3ctag.org/p/10-02-2016-minutes.md

-- 
Ted Guild <ted@w3.org>
W3C Systems Team
http://www.w3.org

Received on Thursday, 11 February 2016 15:33:21 UTC