I also had an arbitrary chat radius, initially. After feedback and testing, I made it so that the radius shrank as more people were online. So, if 1 person was on the site the radius was non-existent - it'd show the next few people regardless of location. As enough people opened the website, the radius would begin to shrink based on your own location to include at least 1 other person as close to you as possible, down to a minimum of like 20 miles. This way, if 10 people were online in San Francisco, but I was building the thing in Dallas, then I could at least join in the discussion in San Francisco until more people in Dallas joined in.
Looking back, to call my old thing a start-up in that post was hilarious. I also learned that my concept, being real-time location-based ephemeral anonymous chat (buzzwords jeez), has a severe chicken and egg problem. It's damn near impossible to reach critical mass for long enough for the chat radiuses to create any opportunity for local discussion. I also learned that it really needed to be a native mobile app, instead of a website. Only tech savvy people really knew what was going on, in regards to enabling location services in browsers.
In the end, I had wished I just made some sort of non-realtime location-based tagging service, instead. Instead of real-time chat, just have location-based notes that you can leave somewhere in a very small precise proximity. That way there's no need for a critical mass of users all at the same time.
Edit: it looks like by the time I posted it to HN even, I had began changing to the non-real-time version. it originally was purely real-time chat. the code is on github: https://github.com/ryancole/geohello.com
This one could have legs. Nearest X number of people gets around the 'this app is a desert' feeling - even if only 100 people are online globally, you've still got a full application!
I feel this results in some obvious paradoxes when considering how the room morphs over time.
One (contrived) paradox: 99 people in the world online in San Francisco and 1 online in London. One more person in San Francisco joins. Do the previous 99 SFers suddenly find they are talking to the new SFer and not the Londoner? Who does the Londoner see they are talking to?
Another (very contrived) paradox: 100 people are logged in who are all located on the edge of a perfect circle. One more person logs in who is located directly at the center of the circle. It's clear who is closest to each of 100 original people (presumably the person on the opposite side of the circle is 'swapped') resulting in each of the 100 original people now talking with the center person, but the center person can not be talking to all 100 people and there is no distance-based criteria for determining which of the 100 to not talk with.
These seem like contrived examples, but as soon as the total user count is greater than the 'nearest X' count they must be addressed. A solution would be fascinating, I don't see a clear one!
I think you might be right on this. If I signed up to a service like this and there was no one in my area, but the app suggested that I might like to chat to these 50 odd people online in London or San Francisco, I might well stick around.
I don't know how you would work it for an app that is supposed to be all about local though. If I sign up, and get connected to, say, some guy called Brian in New York and the conversation is great, but then the app grows really quickly and all of a sudden the app decides Brian isn't close enough for me to talk to anymore, it might leave a sour taste.
There are definitely exciting UX possibilities with this though. I need to think on this a bit further.
sure, but the statement was that you would be talking to 100 people, there is no way to enforce that it is 100 people. Perhaps you could partition the geographic area with centroids like a voronoi diagram, but then you will be dropping people in and out of chat if you move them to different clusters. A fixed radius is the simplest solution and probably a good place to start.
If anyone wants to try this with people, type "Columbia University" as the location. I'll try to stay there all day.
Welcome back, tomd3v! I see you solved the bug where anyone could rooms by URL. Unfortunately, that bug was also the only feature that let anyone use this for chat. Instead of fixing it, you should have spent time on some of the features suggested in that thread. It's already clear you're just going to get the same feedback again this time.
I built something similar. The idea was to be able to play multiplayer games with people nearby. My motivation was simply to enable face to face two player games as the easiest option. The trickiest part, which I never coded up, was to be able to include people who are further away if there is no-one nearby, otherwise the game became unplayable without a cooperating partner.
I think this family of ideas, connecting to nearby people easily, has real legs.
Comes with a built in tankwars multiplayer demo. Alas the URL to the demo no longer works as I shut down the server that was running it. If anyone wants to try it out I would happily get it running again.
Edit: Are you a New Zealander? I am from NZ myself, always keen to hear from Kiwi hackers.
I feel like there is definitely something in this basic idea. Although I'm not sure that a rigid 1 mile boundary really makes a whole lot of sense. I'll explain:
I feel like the experience you get with an app like, say, tinder, which lets a user chat to attractive people locally, is really quite an incredible proposition. In tinder's case the hook is that you are chatting to, and then hopefully meeting with, a slice of the local population that you find attractive and that also find you attractive. You both find each other attractive - so you have that in common.
What I find myself hankering after more and more though is a service that lets you discover a slice of the local population with which you have other things in common. Not a dating app, but rather an app where I could chat about a topic I'm interested in (technology, hockey, model trains, tiddlywinks, you name it) with people who are also interested in that topic and then hopefully connect and meet up with those people who all happen to live fairly locally.
Hacker News for instance is great for tech chat, but as someone who lives a long long way from SV I find that it's much tougher to actually MEET people who share my interest in tech. There are tech events going on around where I live, but it's still an intimidating prospect showing up to some niche interest event when you haven't met or chatted to anyone there before. Technology is big interest of mine, so in the end I did seek out events and other ways to connect to like-minded people. But I have a much less active interest in, say, sailing. It wouldn't normally occur to me to go seek out a sailing club or a sailing meetup of some kind locally, but I might still like to chat about it. And chatting about it with similarly interested people nearby might lead me eventually to go along to such an event and meet new friends and try new experiences.
I feel like Kiwichat goes some way towards implementing that vision, but I just signed in, and my one mile radius shows no chats nearby. I live in Edinburgh, in Scotland, which isn't always the first city to bustle with new technology, but it's maybe 5 or 10 miles in diameter, and I have no way of knowing if someone else in Edinburgh is using this site, but from a location only a couple of miles away. Maybe I'm missing out on some new connection, or experience, all because I couldn't put my radius up to a level that makes sense for me personally.
I feel like this is close to the app I really want. Closer than anything I've seen so far, but still sadly quite a way off. It says on your landing page that it's for "events, meetups, or just chatting with people nearby e.g. students in your campus". If you're at an event or meetup then you are already most of the way there and you are probably talking in person. If you're on a small campus then this might serve the need I'm looking for. But if your campus is big and spread out (like mine was at university), or you live in a small city or town and want to chat to people that you could easily meet, then it's just not serving that need... Yet.
In terms of execution though it looks great, the sign in experience is pretty much flawless. Congrats on your launch.
This is the type of thing where people will show up, see that their room is empty (if you're not just searching for "San francisco") and leave, unfortunately
Cool idea. I started work on something similar, but after some planning I felt like I was basically writing an IRC chat from the ground up and things like tinychat can (kind of) already do this. In the end it seemed like a lot of work for a very iffy project, especially if I wanted it to get popular.
I agree that the 1 mile is too rigid. There are only going to be about 3,000 people within a mile radius of me (~1K/sq mile) whereas in Manhattan you would expect around 210,000 (~70K people/sq mile). To get a similar expectation, I would need a radius of about 8 miles. Though of course, as I expand the radius to the south near NYC the population density gets a bit higher, and as I expand my radius to the north it gets lower and 8 miles is enough to bring about changes.
I see one of these pop up every so often and always get disappointed when there is no one in my area. I wonder if this sort of thing might work better if it connected you with the N closest users? Another option would be to create spatial clusters using something like k-means clustering[1].
Love the idea. Wish the radius was wider. There aren't any chats near me, and I'm in a densely populated area.
Since you're taking people's location, be sure to include a water-tight privacy policy that keeps 3rd parties from accessing location data.
But since you'd have access to the location, it'd make an easy sell for targeted advertising. But please, try to think of other ways to make money. Everyone's sick of ads.
Well to me a way around this could be to leave user bread crumbs. For example if its not an extremely populated area when a user leaves the area, a message could say user "johns" was here 10 mins ago. contact him with a PM or something.
That way you know that users are in the area using the app and may come back.
Without critical mass, this becomes an asynchronous, location-based messaging tool, similar to what foursquare is kind of about. However you could use this tool/tech to geotag areas, spontaneously create chats for events like conferences, or even for emergencies like Mountain View Puma.
It's cute that you're both naming your products after New Zealand's national bird.
Both of your products sound like dating sites dedicated to meeting New Zealanders. That's....a niche market.
Kiwi != kiwifruit, if you talk if eating a nice juicy kiwi in the rest of the world you're either salacious, a cannibal, or undertaking some kind of illegal feast (main course: roast bald eagle perhaps)
The people in my 1-mile radius aren't the same people that are in their 1-mile radius, so everyone would see a fractured group conversation with a bunch of out-of-context responses.
I've been trying the somewhat similar Yik Yak (mobile app) for about two weeks. It had flood of college students home for the summer so I assume it's taken off at colleges.
I was hanging out in Oakland for 45 minutes and didn't see another person come by -- I was expecting the Bay Area would have the densest group of people trying this out.
I also had an arbitrary chat radius, initially. After feedback and testing, I made it so that the radius shrank as more people were online. So, if 1 person was on the site the radius was non-existent - it'd show the next few people regardless of location. As enough people opened the website, the radius would begin to shrink based on your own location to include at least 1 other person as close to you as possible, down to a minimum of like 20 miles. This way, if 10 people were online in San Francisco, but I was building the thing in Dallas, then I could at least join in the discussion in San Francisco until more people in Dallas joined in.
Looking back, to call my old thing a start-up in that post was hilarious. I also learned that my concept, being real-time location-based ephemeral anonymous chat (buzzwords jeez), has a severe chicken and egg problem. It's damn near impossible to reach critical mass for long enough for the chat radiuses to create any opportunity for local discussion. I also learned that it really needed to be a native mobile app, instead of a website. Only tech savvy people really knew what was going on, in regards to enabling location services in browsers.
In the end, I had wished I just made some sort of non-realtime location-based tagging service, instead. Instead of real-time chat, just have location-based notes that you can leave somewhere in a very small precise proximity. That way there's no need for a critical mass of users all at the same time.
Edit: it looks like by the time I posted it to HN even, I had began changing to the non-real-time version. it originally was purely real-time chat. the code is on github: https://github.com/ryancole/geohello.com