Well, it’s a solution looking for a problem.
There is an app on the Google store for Android called Grid Square Locator which gives the eight-digit maidenhead locator, trouble is who outside Amateur Radio users the maidenhead locator system?
OS LOCATE is preferred by Mountain Rescue.
One builders merchants that I use occasionally ask me for my What3Words location for delivery purposes and I have to tell them I don’t have it on my phone and give them directions
Well, sotamaps also gives an eight-character maidenhead locator, but to see it would require the user to click twice in the “Grids” map control to activate the Grid Overlay and then zoom to a particular location in the map, but who in their right mind would be prepared to go to such trouble for such small reward?
Thanks. I was slightly surprised to find that the OS Maps app doesn’t appear to have a similar function.
Hamgps app gives you 8 digit locator plus three dots in the up right corner gives you status of all sensors in your phone
The point is not there are a myriad of position reporting schemes and apps to leverage those schemes, but that What3Words has a big marketing operation pushing how perfect W3W is for emergency responses. It is trying to get people to transition to using its app and thus get commercial licencing fees when it has many, many failings in this scenario.
All position reporting schemes have issues and features. In an emergency you want something that reduces the chances of someone panicked or injured to give a location that cannot be confused. W3W hinges its emergency suitability on the basis that giving numerical lat&long is fraught with errors unlike W3W. That is until you examine the words used and find there are plurals and homophones abundantly scattered throughout the words chosen which fundamentally undermines the very premise of its existence.
It most certainly is, by the vast majority of teams/incident commanders.
From direct personal experience I’d say I’ve had a tiny number of miscommunicated grid refs in over 10 years of incident command. The same cannot be said for W3W.
A few weeks ago, a bandleader gave me a W3W for a gig location. It was a rural residence and the postcode covered a large area. I put the app on my phone and it got me to the gig. It worked perfectly. What’s the issue - that if you slightly miscommunicate the three words, it takes you somewhere completely different? Isn’t that the same for all systems? For instance if I advised that I was in SJ850750 (3 miles from my home in Cheshire) but it was misheard/mistranscribed as SK850750, that would have people looking for me in Lincolnshire.
Surely it’s the standard of communication rather than the system that is the problem?
The ambulance goes to the wrong place and you die. Rocket science it’s not!
The sad thing is, it could have been an excellent system had it been properly peer reviewed. Unfortunately the company now seems more interested in employing lawyers rather than scientists/engineers to maintain its position. Whatever justification is there to include plural and non-plural versions of the same word?
Our local State Emergency Service, where I am a volunteer, has encouraged the use of W3W and I am a bit concerned that too much trust is put into the system. The system no doubt works if used precisely as intended, but as Andy says, a minor change to a word and it has you in another country, or anywhere but where you want to be. So it is still dependent on precise, literal use of the exact words coming from the conversion process. Just as making an error in one digit of a lat/lon or a grid ref, making an error in any letter of any of the 3 words places you in a completely different place.
To me, the fault in the system is that while an experienced operator can tell quickly whether a grid reference or lat/lon is wildly wrong and can ask for a repeat, they cannot do the same with a 3 word code. Any of the three words could be spelled wrongly or miss a letter, and you cannot tell by looking at it, that it is wrong. Add country specific spelling of some words (like Grey/Gray), and you have the makings of an unreliable system.
Incidentally someone asked whether anyone but hams use the Maidenhead locator codes, my Garmin GPS62S has a location code option of MH, which gives a 6 character grid subsquare code.
73 Andrew VK1DA/VK2DA
Which is not my point.
How is it deemed that W3W is less robust for communicating than other systems, such as OS grids? All are capable of a miscopied character. Has it been demonstrated that W3W is less reliable through either mathematical proof or data analysis? All I’m seeing so far is anecdotal evidence.
The unique selling point of w3w is that it is easier to remember 3 words than a random sequence of letters and/or numbers. If you have to start spelling those words out in full to reduce the risk of errors it starts to lose some of that advantage.
If W3W is going to replace more traditional grid references it has to be shown - by mathematical proof if you will - that it is MORE robust for communicating than other systems. Otherwise there is no point to its existance. I think that Andy’s point about the use of plurals and homophones is quite compelling. Numbers and letters in any system can be used or copied erroneously but that use multiplies the chances of mistranscription.
Which is not the point being made Tom. It’s not that W3W is less robust than giving a lat&long or an NGR in the UK, something I specifically said. W3W marketing emphasises how much better is for emergency use as giving three words is easier than say 12 numbers especially when stressed or injured. But W3W words are not unique in the use of plurals and homophones. As Jonathan says, had the words been peer reviewed then the potential confusion points could have been removed before the system was rolled out.
W3W refuses to fix the problem plenty of emergency responders are complaining about because there are plenty of business users with a W3W location on their letter heads, packaging, business comms, websites etc.
An example of the homophone problem:
wants.census.incite and wants.census.insight sound the same, one location is in Egypt and one in Serbia. You only know that once you plug the words into the W3W app and then have to get the spelling confirmed. At which point you may as well have given the lat&long. But at least with a homophone error you get a significant error in location.
An analysis of single character transcription errors shows that
Again, it’s not that it’s worse than other methods or that they are better it’s specifically that it has all the same comms issues other systems. The W3W website says
" Locate incidents faster: When callers struggle to describe their location, what3words helps them say exactly where they are, saving precious time."
Which is true as long as the words chosen cannot be confused and need to be spelled out. But they designed this flaw into the system and are trying to bluff and bluster through the issue.
I am not native english speaker and my accent is wrong and even don t know to read all words (or worlds :-)) right. What to do with w3w?
I will stick to lat/log or mainhed locator