Ask HN: Is building a Google Maps alternative from scratch bad idea? | Hacker News Reader

Ask HN: Is building a Google Maps alternative from scratch bad idea?

want to build a Google Maps alternative from scratch. I’ll start with one city, focusing on the downtown area, and then expand from there. I know there are open-source versions available, but I want to build this map for a specific use case.

On a scale of 1 (a stupid idea) to 10 (a great idea), where would this fall?

View on HN

What’s your specific use case? I would base it off OpenStreetMap as much as possible - building a whole map and keeping it up to date is a lot of work

I vote for openstreetmap, amazing, you can even update it.

It depends on what he wants to do. Creating something like that from scratch is very hard.

If you are building just for yourself, then it's definitely a good idea. Google maps is great for rich POI data but other providers (Mapbox, Here, TomTom) are much cheaper and can be a good starting point to start building your use case fast.

Or go open source route with https://www.geofabrik.de and start with OSM data. Not too hard to setup but more manual than getting an API key.

I am currently trying to build the alternative to Google maps after finding it too expensive (https://www.milaapi.com)

I'm enjoying doing so with https://cartes.app.

Web-only though. Building apps would be crazily expensive.

There is nothing to it. I once made web.py-server that combines maps from two countries. While kayaking. It was too annoying when you had to switch programs to see what is on other side of river. Unfortunately they do follow Google tiles standard anymore.

Originally this program was running on Python For Symbian, which is historic and funny. https://github.com/timonoko/webpy_map_server

Tiles.py seems to be how you decode other systems to the Google-tile system.

I will use C and C++ only, Python too slow, Rust ( I despise the community), JS is poorly designed bloated mess, Java way too verbose, Swift is for Apple only which is limiting, I wanted to use objC but I heard Apple is gonna dump it soon.

C and C++ is the only viable languages here since this project is all about graphics.

Er what? In Python the graphics libraries are written with whatever is the speediest language.

But If you study my program all it does is to produce a website and graphics rendering and viewing is done in a browser.

Map tiles are organized in scrollable array with HTML-div sentences. And map tiles are fetched:

if ny == 'google': return 'https://mts1.google.com/vt/x=%d&y=%d&z=%d' % (x, y, zoom)

There are already dozens. https://gisgeography.com/google-maps-alternatives/ How many millions of dollars do you want to spend on your project before you run out of money?

Do you think it takes millions to cover one downtown ? I am mot competing with Google maps, my project isn't customer facing product, it is a business to business or gov product, the map will be fine even if we don't cover the world or entire state.

> I am not competing with Google maps, my project isn't customer facing

Okay, that sounds much more manageable. But then could you please clarify? Your original question was about an alternative to Google Maps. But now you say you actually want an alternative to a different kind of gis mapping system. If you go into MUCH more detail about the context, we can give you better guidance.

It’s closer to 1 than to 10 (it took Apple with all their billions something like five years to get their maps usable outside of silly valley).

But if you can use OpenStreetMap as a baseline you may have a chance.

Remember my map isn't the world or state or even a city, I will start one downtown couple of blocks, and then slowly build outward.

Apple focused the entire world, which is not what I want.

If you enjoy building it, 10/10 great idea

using GIS?

I looked at getting gis data from my local county. Sent an email and all I got was crickets.

It all depends on where you are at and what's available to tap into, as far as gis is concerned.

I tried to make a tree location map using phone gps meta data a couple of months ago. What a mess, the precision is too low. You can buy precision gps receivers, but they are expensive. It was a great exercise though. Now I know why trail navigation apps fail around closely spaced switchbacks in mountain areas. The gps data is too imprecise.

But maybe you know more than I

Not only gov data, but also custom map I wrote, have you tried Bluetooth beacons ?

You would be surprised how far you can go if you ignore these expensive APIs and build your custom stuff. People are lazy, if any big corp does something, they immeasurably avoid it and start building their 1000th note taker app.

Reply on news.ycombinator.com