What3words, the GPS app that can find anyone anywhere | Hacker News Reader

What3words, the GPS app that can find anyone anywhere

theguardian.com

15 pointscamtarn25 comments

I'm not a huge fan of their service for copyright reasons, they don't publish a database of the words (https://support.what3words.com/hc/en-us/articles/207769875) and the whole thing is copyrighted so you can't reuse it. It makes it pretty much useless as an universal address system for this reason.

Also on the downsides, I don't know for other languages but in French the words are NOT common words and a good half of them would require a dictionary for a native speaker, making the address system useless.

Note the simple words are assigned to populated areas. In the ocean it might give you "overripe.thoroughbreds.salads".

In French, on their map demo which centers the map in Paris, it's already pretty bad. I honestly have no idea what some of the words mean.

Edit: A few examples taken randomly of mostly unused words in French (require a dictionary for a good 90% of native speakers): Volute, redise, dépiler, rapière, ânier, putatif, toison. All of that is in the center of Paris, and it took me 5 mins to find them, there's probably hundreds of rare words.

I honestly have no idea what some of the words mean.

Do you need to? If you don't speak the appropriate language you won't know the meaning of the words for any place. That doesn't make it any less useful so long as you can pronounce them.

The main goal of this project is to have memorable words instead of a meaningless address name, if that's not any better than the existing system, why would we use it?

But you need to know how to write the word and in which grammatical case the word is or otherwise you end up on the wrong end of the earth because every homophone also maps to a real location.

I French this is extremely important.

For one we have plenty of homphones. Is it saut, seau, sôt sceau ?

Then if you do not know a word, say châtière, it can become chattière.

So only homophonically unique words and well known ones should be used

I have no idea how they chose their words, you will get things like the words you listed, say "ânier" which is apparently a donkey-driver, or "chatière" (a cat flap) but their dictionary doesn't even include just "âne" or "chat" (donkey and cat respectively) which are words anyone would know.

This is completely absurd.

Moreover it is absolutely not locality sensitive, rendering the whole system into a proprietary hashing algorithm. Probably the localization is pretty much the only element I agree to be good to have, but I believe it's likely that some word pairs are in the same synset for some languages, even though I'm not really in a position to evaluate its wordlists.

I'm not sure I understand their idea of localization.

The center of Vienna is "decays.jump.graver" (if those are simple words is questionable).

But when I switch to German the same block becomes "fahrende.hügeligen.ansprüche" (driving, hilly, demands). So the words I get by default are again just random strings for someone who doesn't speak German)

And even someone who speaks German may get the inflection of the words wrong and end up in "fahren.hügelig.ansprüche", "fahrend.hügelige.ansprüche" or any other permutation of the many ways the same word could be written in another context.

And worst of all, all those permutations exist and map to other locations somewhere on earth.

You are correct, it's a horrible implementation of an otherwise good idea.

Personally I want a public domain list of 2^k distinct words of distinct and unambiguous enough meanings, that are then translated to many other languages and evaluated for the same criteria (in reality, such system will be required to begin with as many languages as possible). My best guess is that k=10 is possible with a lot of efforts.

> And worst of all, all those permutations exist and map to other locations somewhere on earth.

To be honest they did give a thought about this, as apparently close sounding words are supposed to map to places that are far apart.

It makes sense, as you should at least know if you're looking for a place in Austria or Burkina-Faso, but it also means that if you heard the words wrong then you depend on their algorithm to find a slightly different-sounding address that is the one you were looking for, instead of a more intuitive system where you could just around because if it sounded the same then it should be closer.

And since the word list is proprietary, you really entirely depend on them. Their website shows three different suggestions, the one you typed and two close sounding ones.

Tough luck if you typed "duper.listons.égalisons" and the actual address was "duper.listons.égalisation", you won't get it in the suggestions.

Yeah, I have explored just a bit and seeing addresses like "rotateur.patois.postier" next to "dater.postant.ânier" (... what? ânier?) feels like these things would be very error prone, it almost looks like they gave similar sounding words to close addresses, but actually not.

The words in different languages also aren't translations, they are completely different, which adds a layer of confusion. I really don't see this going anywhere.

This is an advertorial - paid for by W3W.

> Tim Adams travelled to South Africa as a guest of what3words and Gateway Health

In reality, they have a closed, proprietary algorithm. They charge for usage. They are the sole arbiters of what goes on there. They have weak internationalisation.

Hell of a PR team though.

"plus codes" [1] are a far better approach:

- no language involved, only ASCII chars and + signs

- open source algorithm to derive location codes

- works completely offline, as codes are generated on-the-fly

- no central DB of statically defined keywords needed

- no dependency on a company which might go bankrupt at some point or change licensing terms at will

W3W looks to me more like a strategic sell-out - implementing an already existing idea, hype it up, cooperate with some big brands (e.g. Daimler/Mercedes-Benz [2]), sell it for profit…

[1] https://plus.codes/

[2] https://blog.daimler.com/en/2018/03/28/what3words-mercedes-b...

What3words is a proprietary licensed database that should be public/free.

It's a very useful service, and there are some strong arguments why it should be free, but getting people to actually adopt a standard is really hard. It takes a lot of effort, time and money. Without the possibility of revenue it's very likely no one would ever have funded it, so it would never have happened.

It'd have been nice if W3W was the product of a foundation like Mozilla, but it's not. There's no reason to begrudge the founders wanting to turn it in to a business.

This is basically like DNS for latitude/longitude coordinates. Pretty cool if you think how revolutionary DNS was in making the web much more accessible. If this has the same impact then it will be spectacularly successful.

Although it will be interesting to see how they can make money out of it given it is a closed model and how that conflicts with bringing its benefits to poor communities.

I used to find funny that my old school What3words was stress.cheat.mixture[0]. I guess that represents the feeling of studying Telecom Engineering quite right.

0. https://map.what3words.com/stress.cheat.mixture

> “When someone asked where you lived, it would be like trying to remember your wifi router password.” That’s when this idea of three words came up. A bit of maths proved it was possible. “With 40,000 recognisable dictionary words, you have 64tn combinations, and there are 57tn squares.”

Sounds like Diceware.

https://en.wikipedia.org/wiki/Diceware

I think this is a great idea and it sounds remarkably like what DNS servers do for IP addresses. Could we modify DNS servers to provide this information (and potentially better localize it--like how country TLDs are controlled by given countries)? I think could also be very helpful for kids or the elderly in the US--they could remember a phrase much easier than their full address.

The problem with this approach is of course language limitations.

Can we bypass this with a combination of numbers and universal symbols?

Numbers are not universal. In that 8 can be represented as Ⅷ, ב, ၇, 八 etc.

But I like your idea of a mapping system based on emoji :-)

Except not all devices show emojis, and the discrepancy between symbols across operating systems. Not to mention, how do you even explain (say) some of them?

it has been done ages ago, is called the maidenhead locator system, it is better in many ways, for example you can choose the granularity of the grid providing some privacy: https://en.wikipedia.org/wiki/Maidenhead_Locator_System

Reply on news.ycombinator.com