- From: Kis, Zoltan <zoltan.kis@intel.com>
- Date: Mon, 7 Sep 2015 19:51:36 +0300
- To: Mark James <mrj@rbate.com>
- Cc: "public-web-nfc@w3.org" <public-web-nfc@w3.org>
- Message-ID: <CANrNqUfUMqQeGChCq=WtGQ6XZQg3_ycmd+RZgugMQ4X9LbqFKw@mail.gmail.com>
> > Zoltan, whenever my Android tablet taps a tag programmed with a URL, the > default web-browser navigates to that address. So I won't cover the native > app handler option you raised. > > However, that is a policy/choice your Android tablet has made and implemented and need not be the same on all user agents. > Your comments suggest that > > (1) A device receiving from a tag or a peer a Web NFC message with a URL > payload will indeed navigate to this URL, despite the extra Web-NFC record, > and this will occur even if there's an active Web-NFC watch (no message > absorption, always a nav). And > When you have a default underlying platform policy like the above, then yes, the native dispatcher may choose to give it to another handler, fetch the URL, and open it in the browser in a different tab than a listening page expecting Web NFC content. However, the user agent should declare the right NFC intent filters on Android. You'd be right saying that the Web NFC record should then be the first record, and the implementation could use an AAR for the user agent handling Web NFC. But that is thick implementation detail. What boils down to the spec is that it should not mandate the Web NFC record being the last record in a message. Is that where you wanted to go with this example? > > (2) If the page at the payload URL contains a watch that's activated on > page load, that watch will receive some form of the original message. > Correct, if it switches to the same browsing context. Though then it works by side effect. I'd like to see user agents configuring correctly the underlying platform's tag dispatching. Maybe it's worth a note in the spec. Best regards, Zoltan
Received on Monday, 7 September 2015 16:52:06 UTC