location entertainment - where is it?

location based services

Jump to Comments

Location based games and applications SHOULD be a core part of mobile phone usage. In reality they are non existent. To be a bit clearer, I am not talking about LBS (Location Based Services), as in services that are available based on your location - e.g. I want my nearest restaurant or show me where I am on a map. I am talking about social applications, games, community built info applications etc. So what I think should be happening is that location technology should be driving new experiences and creating whole new ways of living - but at the moment location technology is only in some very limited cases helping us do existing things more efficiently - that’s all.

I think there are several reasons for this.

Making any kind of location based service requires adherence to codes of conduct and/or legal obligations. This is of course a good thing - but it does complicate the development process, especially when every entity has a slightly different approach.

When looking to develop a location based app here are 2 options for location applications and games: GPS or cellid.

GPS is really in too much of a niche to develop casual applications even using something like Flash Lite. After all one of the core attractions of location based games/applications is ubiquity - anyone anywhere, not any geek with a GPS module in one pocket and a smartphone in the other. As smartphones such as the N95 start to become available this may change but it’s still a niche market and doesn’t offer the promise of GPS people or friends based interaction.

So the only real option for location based stuff is using the cellid of the phone. There are 3 ways you can use the cellid of the phone to create location applications.

You can do this through an operator api. This involves you calling the operator api over http and getting back location info based on the requested telephone number. However this makes no sense in the ubiquity stakes because it is normally operator specific and therefore only useful to businesses looking to consume business wide services on the same network. I suppose you could do some operator specific stuff - such as games but it doesn’t really connect strongly with the ‘my friends’ side of things.

So then you have to look to a few of the operator location api aggregators who interface with all of the operators within a given territory. So this is great, it offers all the functionality needed and it’s in no way hardware dependent. However because it just feeds off the individual operator apis it is stupidly expensive and therefore not at all conducive to any kind of innovation. I can’t remember the exact cost per hit but its something like 20p per hit and you have to have an expensive monthly rental setup too. So there is no room for experimentation and it rules out any kind of real-time movement application or game where the server is regularly pinged. You could maybe have an application that did a daily request or an on demand request, but this is not what is exciting about location based services - it does not expose the potential that mobile has!

So the third cell id based option (if money is an issue, which it is) is using raw cellid data. This involves using Java or something else to get the cell id. Once you have the number you then have to try and make sense of it. This is the hard part, there is no real logic behind cell ids, they are just a number identifying the base station. Operators don’t publish any valuable information about their base stations (for obvious reasons) so there is no way of matching a cell id to a physical location. The only way to make sense of cell id is via open source initiatives that try and collate useful cell data (few examples below).

http://www.svgopen.org/2005/paperAbstracts/cell-lbs.html
http://www.cellspotting.com/webpages/cellspotting.html
http://gsmloc.org/

The issues with the open data sources are:

they could be open to abuse much more easily than pay for models.
the information is fragmented
the information is frequently inaccurate (not so if its GPS based but if it’s by user input)
any cell data must come from the device over http - so a device must report its cellid over http to a server. This is not the same for an operator because they automatically log this data every time a phone enters a new cell. So even if there were databases full of accurate information there would be the cost of regular data traffic to consider.

So to summarise - what a mess! There needs to be some shifting from the operators. Whilst it’s quite expensive technology to setup and maintain - I think it makes sense that there is a business tier and a consumer tier to pricing. Not only does the business pricing already justify the infrastructure, but the potential mass market of a low price consumer tier would pay for itself and help define the new genre of data applications that operators are so desperate to have. What would be ideal? Well how about a European cross operator aggregator with a unified privacy policy/procedure, they could even aggregate wifi info when that starts to take over