Enabling Location in Applications

momlondon.pngThe theme for last night’s Mobile Monday London was ‘Enabling Location in Applications’. The presenters included Ted Morgan, CEO of Skyhook Wireless, Ben Ward from Yahoo! Fire Eagle, Charles Wiles from Google (Gears), Andrew Scott from Rummble, Justin Davis from Buddy Ping, Mark White from Locatrix and Matt Womer from W3C.

Here’s what I took away from the event (my personal thoughts in italic)…

iPhone 3G actually uses Skyhook. This aggregates cellid, WiFi and GPS information to provide location where GPS alone might not work (indoors, height urban areas). Skyhook already have 50 million wireless access points (16 million in Europe) mapped by driving around various countries. While Skyhook uses cellids, it does not obtain any information from network operators.

If it does all this, I can’t help but wonder why my iPhone 3G is so slow at getting a fix. The maximum time to get a rough fix should be the time to get the cellid (negligible) and do a http or ip based lookup to a server somewhere. This should take seconds not minutes. Also, WiFi and cells change over time – how will this information be kept up to date?

Skyhook also have a ‘loki’ iPhone app that can invoke a request that includes location, to a web site through the web browser. This allows remote web sites to know the user location. It’s a clunky workaround as it requires the user to start an application to invoke the browser in this way.

Yahoo’s FireEagle is essentially a database or broker for current and past locations. Anyone can either provide or consume locations according to end-user defined privacy criteria.

I came a across FireEagle a few weeks ago. I must admit I was sceptical. Why would anyone want to trust Yahoo with their location? What’s in it for Yahoo? Having thought about it some more, an aggregated location information makes sense. The alternative, having lots of applications each providing and consuming locations drains battery. There must be some savings from aggregating this stuff. Also, there are classes of applications (proof of concepts, mashups, hobby) where people don’t really want or need to create their own database of locations.

Google told us ‘the web is the platform’. Interesting given they are championing Android. Anyway, the new Geolocation API will provide a Google Gears interface for mobile web browser based applications. There’s a synchronous getCurrentPosition() call as well an asynchronous watchCurrentPosition() that will allow for more battery friendly applications. The actual location can come from a variety of providers – Skyhook, Vodafone, Google etc. There will be implementations (plug-ins) for Windows Mobile, Android, PC and one other platform by the end of the year.

I am still wondering about this merits of this. I presume the the plug-ins themselves will require the installation of a download which is effectively an application (Ok, a DLL in most cases) and the supported mobile platforms won’t be that pervasive.

The W3C has just started to look into providing a standard location API for web browsers.

It seems to me that many organisations are all doing the same thing (trying to get the browser to support location) but in different ways. Let’s hope they converge otherwise we will have yet another set of fragmented APIs.

Rummble gave an interesting presentation and proclaimed ‘Mobile Internet will be the Internet’. Given enough time, people might use their mobile devices as their main device and hook it up to a monitor at home in place of today’s PC. Interesting. Andrew explained how their first iteration ‘Play txt’ failed because users ‘didn’t get it’. I actually see this a lot – tech people such as us can get too embroiled in things that the (lazy) mass market care little about and projects can go off the rails when mildly complex features detract uptake.

Locatrix’s presentation (and the Q&A) made me think more about privacy. There seems to be contention between allowing the user to control their location privacy and ease of use. The more control you give, the more annoying prompts and options need to be tolerated. Another aspect Mark touched on was white label branding. If you are a startup, consider this as it makes your idea much more sellable.

The most interesting part of the Q&A was when someone asked about the need to know other peoples’ locations without them having to push this information or use power hungry GPS, and the importance of cellid in this respect. The iPhone was briefly mentioned in particular it’s inability to run background 3rd party applications. There was also disagreement as to whether cellid on its own is good enough to determine location.

UPDATE: Interesting read from Nokia Conversations, ‘Location based services still do not measure up? Or do they?’