W3C home > Mailing lists > Public > public-device-apis@w3.org > June 2011

Comments on Network Information API FPWD

From: Jo Rabin <jo@linguafranca.org>
Date: Wed, 8 Jun 2011 05:04:52 +0100
Message-Id: <7BD6B790-3AF4-4061-AA80-57B633951DCC@linguafranca.org>
To: public-device-apis@w3.org
Dear Device APIs and Policy WG

It's good to see progress in this important area.

I wonder, though, what the use cases are for the interface as defined? I'm unclear how you'd usefully employ the interface described - knowing that the device has a 4g connection as opposed to a 3g connection, for example, allows me to do what? It might be worth illustrating this. I do appreciate that there is a level of complexity involved here, for example there may be a series of connection types involved, e.g. a Bluetooth connection to a device that is connected via Ethernet to a router that has some kind of broadband connection, and which it might be useful to know about, but which in turn might be out of scope. But perhaps it would be useful to spell out what the scope is.

Presumably the enumeration of connection types is related to the online or offline status of the device's current connection, rather than the device's capability, since only a single value is returned - and therefore "none" means the device is offline, but could be online? It might be worth spelling this out.  Likewise if the current connection is 2g then an author might, before a large download, recommend switching to WiFi, but would find it useful to know that a device does not have a WiFi capability before prompting to make such a connection?

Is it worth considering the case where more than one connection is in operation? Would this imply some kind of mechanism to control the connection type too?

I am concerned that the current extensibility mechanism defined for the values will lead to interoperability problems.

What is the intention in respect of how this interface relates to the Network interface of the SystemInfo API, which appears to offer both a broader vocabulary of connection types and a wider range of possibly more useful information concerning bandwidth and likely costs (via the roaming attribute)?

I'm curious to understand how the choice of connection types was made, for example why was the choice made not to distinguish different types of possible WiFi connection? Would it be useful to define more clearly what the meaning of the labels 2g 3g 4g is? Would it be useful to distinguish (wired) Ethernet connections at 10Mbps 100Mbps and 1Gbps. 

In respect of the issue identified under section 4 - without knowing the nature of the debate around this, I'd offer the view that if it's useful to be able to identify the connection type then it's useful to be informed of a change in status of that connection type.

Hope these comments are useful in some way.

Jo
Received on Wednesday, 8 June 2011 04:05:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:14:21 GMT