Ken Ng : It Takes Time To Get A Map App Done Right

To get to the real KL International Airport (KLIA), you actually have to type LCCT!

LCCT is the adjacent Low Cost Carrier Terminal, but not according to Apple Maps!

The thing about getting mapping right due to the scale of data is to get it out to be used by everyone.

The issue lies at the fact that (in my personal opinion) Apple isn't really good at dealing with software that leverages off huge, largely unstructured dataset. Look at Siri. They tried to set the customer expectations by saying it's a beta software, probably knowing very well that the initial dataset that they may have to accurately recognise speech is far from sufficient to do the job of recognising pretty much everyone who would be using Siri. And the only way to get Siri to work well is to have an immense database of data to process, and guess where the source of it will be from? Us the users of course! The more you use it, the better it will become. But it's just against Apple's perfectionist nature who would usually only release software that works well and within their control. They simply does not seem to know how to work with their users to help them get things right.

I think Maps is just the same. What they've done in Map 1.0 is to clobber together a whole bunch of POIs from various data sources, including TomTom's. The result unfortunately is really messy. Why? First, you have the understand how complex getting these data right are. There's simply no way software algorithm can do it and it is also absolutely impossible for a small maps team to vet through all the data to get it right. And unfortunately for Apple, TomTom's data set is very lacking to begin with, so as right as those data may be, the lack of data allowed sub-standard data sets to proliferate in the maps app, perhaps even screwing up TomTom's data set itself from the merging algorithm used by Apple.

When you deal with POI data, things aren't as simple as 0s and 1s. Simplistically, a POI metadata would consist of at the very least, the name of the POI, it's category, and it's geographical location in longitude and latitude format. Here's why merging those data isn't as simple as it seems. Let's take a fictional example of a imaginary Burger King POI in the land of Timbaktu. Let's say one data source has it stored as "Name: Burger King", "Category: Fast Food", "Coordinate: 38 53 23N , 77 00 27W", and another data source has it stored as "Name: Burger King Restaurant", "Category: Restaurant", "Coordinate: 38 53 25N , 77 00 28W". By pure comparisons, these are NOT the same data. So you need to use complex matching algorithms to programatically determine if those two data are the same or really just two somewhat similar named POIs that's located near each other. But to a human, we would instantly relate those two as the same Burger King outlet, and especially so if you are local to that area. And this example is far more simplistic that what Apple has to actually deal with.

So in order to really get things right, not just in popular areas but pretty much every corner of the world, Apple really do need everyone's help to chip in and correct the data in Maps by having user report issues and errors they find using the "Report a problem" option in the app. My point is, they've got to start somewhere, and they did. And in my opinion, it's hell a lot better than Google Maps 1.0 especially for us here outside the US. I remember that my condominium here in Malaysia didn't even exist in Google maps until a few years after it was completed. And Google themselves took many years and a lot of user feedback to get where they are today. Sure, they've spent perhaps billions in acquiring satellite images and topography maps. Apple can and have bought the same imageries too. However for the POIs, location metadata that is the main challenge and while they can try to buy those data, I believe that local POIs can only be obtain from the local users themselves. Crowdsourcing is really the only true way.

I think Apple has the means to get it done. It's popular enough worldwide to have local user provide the data. So I think that all they need to work on now are their communication skills in dealing us users to be part of the team to get Maps right.
The other problem is that Apple simply does not have enough people with the expertise to deal with a mapping solution. Apple really need cartography experts, a lot of them if they want to get this right, fast. Perhaps Apple would just go out and buy someone to fill up the gap, Garmin maybe? TomTom's dataset is a joke to be honest compared to mammoths like Google Maps and Nokia's Navteq. And how big is Google's Map team? Last I read, it was at 1,100 full time staff and 6000 contractors worldwide. that's slightly more than 10% of the entire Apple workforce.

You may perhaps ask, don't Apple do some field testing at all? I really don't know. Perhaps that's the problem. Apple is simply just too secretive in field testing prototypes and unreleased products that perhaps it never had the chance of being extensively field tested at all, and thus contributing to this whole bunch of weird POIs outside of the Cupertino and popular spots that the testing team would have checked.

Sure, the end result is bad. But is it a total disaster? I think it's taken out of context just because it's "Apple the Goliath". I've used the new maps app on iOS 6 and so far it has been good enough to bring me where I want to go. And it sure is better than Waze which everyone seems to point out as an alternative to #iLost maps. Lastly, a mapping project at this scale is not an easy project to undertake, not even for an organization the size of Google which took years to perfect. But hey, I'm glad that an alternative to Google and Nokia's Navteq is being build. And really, who else can pull it off today when you need not just the cash but also the will and means to continuously fund this project without any ROI for years to come.

Here's a challenge to you. Sure, it's your right to complain, but what are you doing to get things right? Reporting the problem merely takes a few minutes of your time, so why not make the effort to help get this right? Stop just complaining and do something about it for a change.

Support Tech ARP!

If you like our work, you can help support out work by visiting our sponsors, participate in the Tech ARP Forums, or even donate to our fund. Any help you can render is greatly appreciated!