Everything is easier with a three-word address

The service does not support Google Maps, but can be embedded into other services or accessed via the what3words free map tool.
What they need is for the likes of Google Maps to support their three word results. Until that happens this aint.going.nowhere
 
What they need is for the likes of Google Maps to support their three word results. Until that happens this aint.going.nowhere

LOL!

I don't usually lurk on reddit, but there are some funny comments: https://www.reddit.com/r/InternetIs.../specify_a_location_accurate_to_3_x_3_meters/

FWilly 25 points 2 years ago
Exactly! Now you see how excellent and useful this is. You can also easily see that it is exactly four cells away from aint.gonna.happen and six cells away from get.bent.twats /s
But they'll keep spamming Reddit with this crap and people will keep saying; 'What a great idea!', because they're morons.
I'll just stick with good old X,Y,Z coordinates like Latitude, Longitude, and Altitude, thanks.
 
Holy moley Jan, that is old news indeed.

Years ago I posted a thread about the better non-proprietary open source alternative: http://www.mapcode.com

I regularly use mapcodes whenever I navigate somewhere with my TomTom GPS: mapcodes are supported by TomTom, so I just convert the GPS coordinates of exactly where I want to go into a mapcode and then plug that into my GPS (I find regular addresses are a PITA due to their length and incomplete nature).
 
Holy moley Jan, that is old news indeed.

Years ago I posted a thread about the better non-proprietary open source alternative: http://www.mapcode.com

I regularly use mapcodes whenever I navigate somewhere with my TomTom GPS: mapcodes are supported by TomTom, so I just convert the GPS coordinates of exactly where I want to go into a mapcode and then plug that into my GPS (I find regular addresses are a PITA due to their length and incomplete nature).

Thanks, makes more sense, and is a better algorithm IMO. See: https://en.wikipedia.org/wiki/MapCode

P.S. It was just over a year ago that this thread was posted in IT Discussions >> Software: https://mybroadband.co.za/vb/showthread.php/866158-Developers-are-you-aware-of-mapcodes
 
Last edited:
In all fairness, I realised this story is just a paid promotion, so jah ... moving along. Ironically a year later NOTHING has changed on the address front, and I'm going to lay the responsibility for that with developers. If you're developing anything that uses addresses, it is zero extra effort and cost to add the open source mapcode APIs. Give it to the marketing team to do some consumer awareness, and we'll finally move towards fixing South Africa's growing addressing crisis. Not doing it with all the tools available for free is just laziness.
 
An example of a location is the Virgin Active in Rosebank, which is “breathed.urgent.fleet”.
"I said breathed you morons, not breath. Now come and deliver my package immediately!"
 
Some interesting alternatives (including What3Words) in this BBC article: http://www.bbc.com/news/business-43544007

Russian start-up Naviaddress is one of several companies trying to solve this issue using digital wizardry.

But it claims to be the first employing the blockchain as way of verifying and storing its digital addresses in a decentralised ledger spread across thousands of computers.

The firm has come up with a way of encoding lots of useful information - photos, route directions, geographical co-ordinates, opening hours - along with traditional physical address details, in a short sequence of digits that functions as a "smart address".

"We call them Naviaddresses - unified digital IDs for any place and object in real and virtual worlds," says chief executive Dmitri Moiseev.

"It will provide people and businesses with the freedom and ability to create, buy, own, lease and sell their digital addresses within several minutes."

Russian delivery company ExpressRMS has begun using Naviaddress to help improve its service.

"Now, instead of the address in Cyrillic, the client receives a short message with a digital address containing the detailed 'last-mile' navigation information," explains ExpressRMS boss Konstantin Yakunin.

"You just follow the 'Naviaddress 1234' link and receive all the information about the location of the warehouse, like address, gate and building number, pictures of the place, and description of route," he says.

The new system has reduced the number of calls asking for directions to warehouses "by 80%-to-85%", says Mr Yakunin.

Naviaddress, which numbers DPD, Amazon and Booking.com among its clients, says it has created 1.5 million smart addresses so far, 1.3 million of those for businesses, and it has more than 60,000 registered app users.

"It takes three phone calls to get food delivered to your door in Nairobi, which is super-frustrating," says Timbo Drayson, founder of Kenyan company OkHi.

"To get someone to your office for a meeting, it takes multiple phone calls, which can be detrimental to the business when you're busy."

Mr Drayson, who used to work for Google, founded OkHi to address the issue.

"We discovered that GPS by itself was not good enough to get someone reliably to someone's door in urban, slum and rural areas, so we invented an OkHi address, which combines GPS, photo and phone number," he says.

Sagi Vaknin, managing director of Nairobi-based restaurant Artcaffé, says using OkHi has markedly improved food delivery times.

"Prior to OkHi, a rider would make three calls on average to the customer to get help with guidance to his location and take at least 50 minutes to deliver the food," he says.

For example, Mercedes A-Class drivers can get in the car, speak three words, and be guided to that 3m x 3m position without any typing or other interaction required, What3Words says. it is working with other car makers, too.

Digital mapping should improve with the advent of the internet of things, believs Peter Kurpick, chief platform officer at Netherlands-based data mapping company HERE Technologies.

The world is "on the cusp of an explosion in location data", he says.

Some forecasters are predicting that by 2020, there will be 200 billion connected devices whose data could be harnessed for location services on a much more accurate scale.

"The problem is that the current development process for location-centric services and applications is rather lengthy and complicated," says Mr Kurpick.

"Data is often in silos and incompatible, while big data management and processing are complex and costly."
 
Gordon, I don't think we should confuse address shorteners with address enhancers (which the second solution is), or Location Based Service integration (which the third example is) or just general all-purpose LBS big data fluff that sounds impressive but actually isn't news at all (data is only useful if it is accessible - I could have told you that back when I was still using floppy discs).

The first example, apart from trying to drop 'blockchain' into the mix to jump onto the bitcoin bandwagon (belatedly), seems like an overtly commercial approach. "You give me all your business' information - like the yellow pages, which is reallllly innovative - and I'll sell you a shortcode for that, and then if your clients or staff give me that shortcode, I'll return alllll the information to them." That's not revolutionary, is it?

The problem with systems like that (essentially the what3words model) is that the system is only as good as the general public are *aware* of the code and what to do with it in the first place; and the 'enhanced' systems that return more than a lat-long coordinate (e.g. with photos, contact details etc) require an internet connection to call up that data. In other words, the enhancements aren't baked into the code itself. Inherently that makes the system vulnerable (what happens if my clients don't have internet access, or the information gets outdated, or there is a storage failure at my host, or the service provider goes out of business). You go and bake something like that into all your marketing and business processes, you're going to be in for a fun time when something breaks.

All the above said, I'm happy that more businesses are *trying* to innovate in this space, but I think we all need to be more aware of what the opportunities in this space are, and what avenues should be pursued and which should be avoided. We need a future without 'formal' free-form addresses, but hopefully that future isn't one where all the addressing systems are competing, failing, being replaced, ported across and with massive barriers to entry (sound like anything people already know?).
 
Gordon, just for you and anybody else interested in this stuff ...

We all recognise that formal 'conventional' addresses are bad, right? In the best cases you can't see the house numbers or street names, in the worst cases you don't read or hear the address correctly and just get completely lost, or in a lot of cases there are places that just don't have formal addresses (e.g. informal dwellings, places in the middle of a field with no roads nearby etc).

GPS coordinates make this simpler, in that the GPS system doesn't care about the reality on the ground. Thanks to satellites, the GPS system can link any position to a specific latitude and longitude ... and you are where the two lines cross. The problem with GPS coordinates is that they're also hopelessly complex and one can easily make a mistake while writing them down, or sharing them with somebody else.

Address shortening systems try to solve this problem by taking a complex lat-long for a specific point of interest on a map, and then converting it into something easier for humans to understand, remember and share.

Now there are two possible approaches to address shortening:
1) The shortcode doesn't mean anything by itself, but is simply a variable that is assigned a complex set of values. It's like if your name is Richard Peter Julius the Third, and you ask me to call you Pete. "Pete" is a short-code for your name, but there is no way I can work backwards from only knowing "Pete" to what your name is unless you've told me what it is, or there is somebody I can ask who you've also told previously.
2) The shortcode itself has actual meaning, because it isn't just a random variable but part of a system that has encoded the lat-long coordinates in a different way to make them shorter, and can be similarly decoded (so long as you know the system).

What3Words and Mapcodes are both versions of the latter, because their founders have essentially divided the earth into a grid and assigned different codes and sub-codes to each block in the grid, and so long as you have the software decoder that knows the full grid and all those codes, it can then convert the codes back and forth into conventional lat-long addresses (and when this electronic conversion is integrated into software the whole thing becomes seamless because almost everything accepts conventional lat-longs for addressing).

The difference between What3Words and Mapcodes is that the first is proprietary, and the second is open source. What3Words also requires 3 words, while Mapcodes are typically this length: NLD 49.4V (the first three letters are a standard country code - so not really part of the complexity). Either system has the same accuracy of GPS coordinates, and are easier to remember for humans, so it's a big step forward.

Now the challenge is getting the developers to keep integrating these converters into their systems, and hopefully settling on a single standard. E-mail addresses work because there is a single standard format for how an address should look, so they're instantly recognisable ... we're still living through the VHS vs Betamax fight of addressing systems.

I find it all fascinating, just because of how important addresses are to every facet of our lives ... from whether you can navigate yourself around, to whether your service providers and emergency services can locate you and the others they need to, to even intangibles like how confident we all are when going out into the world and venturing into unknown areas.
 
Last edited:
Conventional addresses work because they are descriptive. They may not be instantly findable but the reason they work is because there's no single standard format so anything can be described. Any other system is fixed so assumes some standard conditions. But what if you're in an apartment block how will you know if it's the 5th floor or the 1st? What if the system isn't accurate enough and overlaps two addresses? Anything that doesn't have an option for adding extra non-standard information isn't complete.
 
Conventional addresses work because they are descriptive. They may not be instantly findable but the reason they work is because there's no single standard format so anything can be described. Any other system is fixed so assumes some standard conditions. But what if you're in an apartment block how will you know if it's the 5th floor or the 1st? What if the system isn't accurate enough and overlaps two addresses? Anything that doesn't have an option for adding extra non-standard information isn't complete.

Any system will get you only that far, but I think we can all agree that conventional addresses are:
- less accurate (a street number doesn't necessarily guide you to the entrance to use)
- more prone to being misheard (simply more information to transmit and receive by humans)
- completely unable to guide people wherever there aren't roads (or even where there are roads but no conventional road numbering - like along rural streets).

So while getting non-standard information to end-users (like photographs, text descriptions etc) will be good, I think the base addressing system still needs to be standardised for uniformity and compatibility. How about adopting a hybrid standard, where the address shortener gets you to the front door or public parking of an apartment block, and then a few letters navigate you within the building (e.g. 4FA12 -> 4th floor, apartment 12).
 
What they need is for the likes of Google Maps to support their three word results. Until that happens this aint.going.nowhere

You can leverage their engine with google maps. Wrote a tool for our company that does exactly that, geocoding in both directions and words.
 
Pity post codes aren't as specific here as they are in UK.
You can leverage their engine with google maps. Wrote a tool for our company that does exactly that, geocoding in both directions and words.
Writing a tool isn't an option for most. Google, et al, need to do it.
 
Top
Sign up to the MyBroadband newsletter
X