W3C home > Mailing lists > Public > public-web-and-tv@w3.org > April 2014

Performance Measurements of Web APIs on TV-like devices

From: Yosuke Funahashi <yosuke@funahashi.cc>
Date: Wed, 30 Apr 2014 21:14:33 +0900
Message-ID: <5360E929.1000905@funahashi.cc>
To: "public-web-and-tv@w3.org" <public-web-and-tv@w3.org>
Hi all,

Regarding performance measurement of Web APIs on TV-like devices, the Hybridcast 
2014 Project, implementing apps on HTML5 browsers embedded in TV sets on the 
market, reported the importance of performance issues at the Munich workshop: 
tweaking animation widgets to achieve acceptable or better UX was the most time 
consuming part of their app development process, especially choosing Web API to 
rendering 2D animations for their apps needed lots of try-and-errors over 
multiple combinations of browsers and devices. This can be a major factor to 
increase their production cost, thus can be an obstacle to expand open or 
device-independent app market on Hyrbrid TV. [1]

I think this is something worth we tackle in W3C and share the result across 
different regions, SDOs, and/or, industries. Assuming we could agree on this, I 
would like to give you some thoughts about how we proceed this project: 
deliverables to be achieved and human resources to pursue them.

Here is a straw list of potential deliverables:

1. A note on best practice about performance measurements and issues
    on Web APIs on TV-like devices
2. A note about existing benchmark software useful for the
    broadcasting or TV industry
3. New benchmark software
4. A proposal for performance measurement APIs

I for one prefer small start or tracer-bullet way of project management: we 
concentrate on the deliverable #1 at first, and expand its project scope after 
we learned lessons from the #1 process.

Here is a draft list of candidate W3C groups where this topic can be discussed:

1. The IG
2. Web Performance WG
3. Other existing W3C groups
4. New CG

Currently, Japanese TV broadcasters have shown their interest to make 
contributions on this topic. If other members of the IG or the IG public mailing 
list are interested in this topic, I think the IG is the best place to deal with 
this topic. Otherwise, directly jumping into Web-perf WG or finding other 
interested parties from game or mobile industries would be better.

These are just my straw ideas. We can talk about how we deal with this topic in 
the IG calls.

FYI, I'd like to provide you with some background information on this topic.

* Timing control for script-based animations [2]

This is a work of Web-perf WG which is related to performance issues of 
animation on UAs.

* 2D performance issue for mobile web apps with benchmark analysis

Here is a result of 2D rendering Web API performance measured with a benchmark 
software [3]. Note that the 'default' browsers on each platform were used to 
measure this benchmark.

|               | iOS  | Android2.x | Android4.x          |
| Canvas        | Good | OK         | Bad                 |
| DOM Animation | Good | Bad        | Good                |
| WebGL         | N/A  | N/A        | Partially supported |

Best regards,

[1] The 1st Implementation Report on TV Programs on Hybrid Broadcasting
     System using HTML5 Browser -Hybridcast 2014 Project-
[2] http://www.w3.org/TR/animation-timing/
[3] http://suemasa.me/test/

Yosuke Funahashi
co-Chair, W3C Web and TV IG
Chair, W3C Web and Broadcasting BG
Researcher, Keio Research Institute at SFC
Special Adviser, Tomo-Digi Corporation
Received on Wednesday, 30 April 2014 12:23:21 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:44:13 UTC