Hacker News new | past | comments | ask | show | jobs | submit login

We wanted to create a routing app for bikes, which would take in consideration cliffs, dangerous streets and the likes.

With our internal team, we’ve managed to create a killer backend that would do all the hard routing stuff without issues. It was better than Google Maps in most cases!

Now, we just needed the app. We used all our money to pay a third party contractor to make the app, and they failed miserably. We never received a working prototype. We ran out of gas.

The backend still exists, in case anyone wants to give the app a shot. I might release the backend as open source. Ping me if this interests you.




I didn't quite understand this. Presumably, you're developers and just created a I assume fairly complex backend product (better than Google Maps). What stopped even just one of you from picking up a book on react-native/electron/whatever and knocking out a workable front-end? Did you exceed your runway, run out of time etc? edit: to clarify, I meant after the third-party didn't work out. Also not critiquing just genuinely curious


I simplified the story a little bit, but two factors contributed to the sad outcome:

1. making a _live routing_ app is tougher than it looks

2. our team was really emotionally exhausted after this (the whole process consumed around 2 years) and some key team members needed some time off

But yeah, it could have been possible to make the app ourselves.


I hear you, and appreciate your honesty. Not many founders talk a lot about the technical challenges of building the app itself. A lot of startup stories I read take this as a given, whereas I'm sure it's one of the biggest factors if not the biggest that causes many startups to fail.


Uber advanced technologies, routing tech lead here. This sounds interesting. Lets chat?

You can email me: michaelv at uber


> We used all our money to pay a third party contractor to make the app, and they failed miserably

Where did you hire/find the third party contractor?

Friends of mine, in two separate startups, got burnt badly by those outfits that offer the "native apps for iOS and Android" package for $10K or so.

I talked to some of those "developers" once, when I helped my friend troubleshoot an issue. It was horrible. They were so inept, they couldn't solve a problem if you provided a solution to them (literally -- they failed at copy-pasting).


And then they want you to sign a document or email saying that they did all they would say and just rip people off when they think, "oh an iOS app with backend for $5000".


We found them in Google. They were from the same country as us. They were really bad. One of the versions of their _navigation_ app would not navigate - the GPS simply wouldn't track the user in the app's map. It was a really frustrating experience.


Hey, we're working in the space of navigation by bike as well, though not in the exact same way as your startup from what I understand. It would be awesome if you could share your backend with us (email: rohan.pandey at gmail). Thanks!


Yeah, frontend work is routinely under estimated. You are not the first.


Yeah. It's not just frontend, you need to figure out the UX too. Backend is usually easy compared to this.


The backend was damn hard. But so was the app! Routing and navigation is a tough domain.


I've been working recently on something semi-similar for runners - calculating elevation gain, difficulty, route generation, etc. First time around I was using Mapbox API but ran into some difficulties and have since switch to Google Maps which still isn't exactly what I'm looking for. Would be interested in learning more about your backend if you don't mind sharing - kguebert93 @ gmail.com


You should write up a Medium article with some test results showing off the backend. You may be able to get someone to buy it before going open source.


I would love to see the backend open sourced as well, as I think this is something that could be quite unique to each country. I'm in Melbourne, Australia and biking safety is a big deal here - i ride everywhere. The uniqueness of each area would lend itself well to localisation via open source. I'm at zubin.pratap at gmail - do let me know!


I’m trying to squeeze in some time to do that. Will make sure to ping you if it happens.


That sounds really impressive. Definitely interested in seeing what you've got (and an open source release would be cool)!


This sounds interesting to me (the releasing as open source bit, in particular).

What's the backend tech stack?


A combination of PostgreSQL + PostGIS + pgRouting, Java + GraphHopper and my personal favorite: PostgREST!


Hey haolez, glad to hear you like PostgREST :)(I'm the current maintainer).

If you manage to open source your project please stop by our docs repo, would be great to have a real world PostGIS project in our ecosystem list.


Will do! Congratulations, btw. PostgREST is incredibly stable :)


Sounds like the folks at https://osgeo.org might be interested...


Please release this, the OpenStreet Maps community could really benefit from a better routing system




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: