Locations are hard

Suzanne Hillman
3 min readJan 20, 2017

Turns out that figuring out people’s locations is hard, especially if you want to try to reduce the amount of work someone has to do or if they are likely to be using a mobile phone.

For some reason, I’d thought that this was already a solved problem, so was somewhat surprised when feedback on a mockup made me question that assumption. After pinging Máirín Duffy to find out if we had access to a database of countries, and how they break down into cities/states/provinces/etc, she realized that we needed a longer discussion.

One thing she wondered was if we actually need street addresses. Given that the goal is to be help people find each other nearby, city was almost certainly sufficient for that purpose.

In many cases, especially on mobile, we would have access to GPS information. So, in that case, we can just show them where we think they are — with a map on which we overlay the city and country information and in which we will make sure that that information is appropriately accessible — and they can adjust as necessary.

On computers, we may have access to the information provided by the web browser, and from that we can similarly show them a map with their city and country information. In this case, we may end up being wildly inaccurate due to people using VPN connections.

In both mobile and computer cases, people may not want to share that level of detail. So, for this case, we would use their IP address information to guess, and display the same thing.

Finally, it is entirely possible that a connection error would prevent us from actually having location information. In that case, we would show a zoomed out map on a computer, with empty city and country fields. On mobile or if we cannot tell where they are, a blank map, with empty city and country fields.

In all cases, people can edit the country and city information to make it more accurate. The ‘city’ field will offer type-ahead suggestions which will include any division between city and country that is relevant. For example, if someone is detected as being in Buffalo, NY, but is actually in Boston, MA, we would offer then Boston, NY first, due to proximity, but also show Boston, MA. And anyone can continue typing to get more specificity, or select from a list of visible options. If, however, the country field is incorrect, they will need to change that before the city suggestions will be correct. As with the map location information, type-ahead suggestions need to be appropriately accessible to people who cannot use a mouse or cannot see the suggestions.

The problem with the type-ahead suggestions is that we still need access to a database which contains that information for each country. There are a couple of options, but that problem remains to be solved, and is a large part of making location information actually workable.

This was an unexpectedly complicated discussion, but I’m very glad we had it. For more information, please see issue #286.

Previous, affinity mapping and brainstorming.

Next, stop worrying about the world and work on internship.

--

--

Suzanne Hillman

UX Designer (https://suzannehillman.com/portfolio.html) who was previously a Linux quality assurance engineer, unofficial doc writer, and a psych grad student.