Infrequently Noted

Alex Russell on browsers, standards, and the process of progress.

Comments for "Mobile Ajax" Slides

i've been pushing for allowing for mobile compatability testing for a few of my projects with ULTRA16 recently. i used to work within Nokia's wap emulator years ago, and the idea of a contemporary mobile emulation environment sounds really good to me!
Would have tracked back but I don't have them on my blog. I don't know if you read my blog (you should. :-), so he's notice of my bit of evangelizing:

I hope you mentioned Trolltech's green phone too.

I was unable to download the slides. I am getting an object not found message.
by Ash at
Interesting presentation. I would definitely agree that latency is the biggest issue with the mobile web. Even with HSPDA bandwidth, it's still a problem. The setup time and amount of hops it has to go through just to connect to a web server is quite large. From an implementation perspective, the other issue is that building a complete DOM on a constrained device has its issues. I would like to hear what your "ideal" mobile browser/platform feature set would be to support the kind of apps you envision? (leaving out any influence carrier's have on the matter of course :-)
by Jeff at
Hi Jeff,

I covered it a bit in the talk, but I think that mobile browsers need to start adding:

  • a form element type that will (with user permission) include location (cell tower ID or GPS coords) in a POST or GET
  • form elements for capturing camera and video data, either from the devices or from internal storage. Normal file upload elements might work, but I can imagine situations where you'd want to specify a media type to filter by
  • audio capture (with similar interface as video/picture)ways to mark part of the DOM "for mobile devices". CSS kinda, sorta, gets there, but it's a hack and mobile devices (rightly) ignore it much of the time
As for the "full DOM on limited hardware" problem, I empathize, but frankly Moore's Law is going to solve that sooner or later.


by alex at
Sounds like flickr geotagging :-) Would a script approach to the location information be acceptable in your mind? GPS performance can be slow on converged devices so it would probably need to be an asynchronous callback unless it was using the last position update.Secondly at my employer, we are struggling with the exact approach to take. Should the main goal be to support current Ajaxified websites on a mobile UI or to support building mobile sites using Ajax techniques? From our perspective the former is a worthy goal although much more complex.One last thing that I'm seeing is that with multi-mode devices (cellular+WLAN+?) device latency is improved in hotspots/home, which is driving the need for better mobile browsing experience.
by Jeff at
Hey Jeff:

Yeah, geotagging of all user inputs is an obvious application. Others, like "location based search" are less obvious, but I think no less important. I understand that it takes a long time/good antenna/lots of cpu to get "5 stars" for GPS acquisition. As part of a form post in a web page, I think less-accurate but quicker to acquire location data is probably going to be most broadly usef. If users need to provide better accuracy, perhaps there can be a UI that will allow them to explicitly choose to get/give GPS coords (and keep them informed of their star-lock count).

I think the main goal here is to reduce the impedance mismatch between the inputs that mobile phones can easily provide and what the traditional web has grown up expecting. I'm not entirely sanguine about the utility of Ajax in this least not yet. Ajax is an upgrade to apps that are already useful. We're still struggling to make web apps useful in the mobile context. In many ways, I think we'll see Ajax as being orthogonal to the raw utility of the UIs that will get delivered to the devices. In some cases, JavaScript will probably be used as a tactic by page developers to fix UIs where browsers don't, but for Ajax itself to be useful, there has to be a significant net win in responsiveness for using it. We'll need significantly more dense screens before I think that'll be the case. I give Ajax on phones probably 3 or 4 years before people will start to do useful things with it. Improvements to what mobile browser can output (location, video, audio) however will pay off in a much shorter timeframe.


by alex at