Comment by pona-a | Hacker News Reader

Comment by pona-a | Hacker News Reader

Parent Full thread

"Nobody uses"...

Do you understand what you just said? Raw GPS is universally unusable, requiring up to half an hour of waving your phone under clear sky, hoping it a) will actually work; b) will not loose signal; c) will not drain the battery before you reach your destination.

The only way to get reasonable estimation of location? Location services. Guess who are the only providers? About 5 or something companies, all with expensive licenses and weird terms of use.

And here comes Mozilla, providing this same expensive exclusive service for free as a convenient API. Just send some WiFi SSIDs and signal strength, cell towers, and Bluetooth devices, and without even having to turn on the actual GPS, you get a faster, more accurate estimate of geolocation, down to a few meters.

This service is vital to mobile Linux or even deGoogled Android or non-Chrome browsers. In its time Canonical had to plead with Here maps for years to give their devices proper GPS. To say a project unique in its entire domain on which hinges the existence of entire ecosystems is less important than a yet VPN or email alias service seems downright silly.

Besides, it's not like Mozilla did it on their own volition. A patent troll had come to destroy their competitor on shaky legal footing, knowing they won't fight back...

What year did you last use GPS, it’s even usable deep within cities now. Modern devices use GPS, GLONASS, GALILEO, and various regional ones too. Locking on is super fast and it’s hard to lose unless you walk inside of a big building. I can even get a good GPS lock inside my house, it’s really good these days.

You’re right that it can use more power, but it’s really not that bad with modern hardware.

With an up to date AGPS database, GPS is quite usable. AGPS is also distributed as a service, though, often by companies that made the GPS receiver chips. If that were to fall away, you'd struggle to make GPS usable for phones.

You may not need Google for a GPS lock, but you certainly can't get a good fix without an external party. I've unwittingly tried this, when I saw a weird HTTP URL on my network being polled in the background; I blocked it in my Pihole and found out not much later that my phone's GPS started taking longer and longer to get a fix. Not "waving your phone 30 minutes at a clear sky" long, but definitely to the point where using Google Maps with cell/WiFi scanning disabled become very annoying.

I run a modest fleet of mostly-offline data collection devices that rely on GPS; we pay a lot of attention to how long it takes before they can sync. With a good sky view, the GPS chip often has a 3D position lock before Linux has finished booting.

What you're seeing is the effect of AGPS and the interaction with the phone's aggressive power management. https://en.m.wikipedia.org/wiki/Assisted_GNSS

It's not really a limitation of GPS, it's about using the external service to let you minimize the time your GPS receiver is active. You can get a great fix without AGPS, but you can't both keep the GPS off 99.9% of the time _and_ get a fast, great position fix.

> You can get a great fix without AGPS

Yes, after waiting minutes with good line of sight for almanac and ephemerides to download from the satellites. Anything that gets a fix faster is already "assisted", usually either by fresh data downloaded online or not-yet-fully-out-of-date data stored locally.

Sure. I was simply observing that this phrasing:

> but you certainly can't get a good fix without an external party.

is not true. You can get a great fix without a third party. You just have to be more patient for your TTFF.

* Unless you want RTK-level accuracy, in which case you're back to needing a third party.

I've been working on a phone that didn't have well-integrated AGNSS and used MLS for WiFi and cellular-based location when it didn't have GNSS fix. I'm perfectly aware that what you're saying is technically true, but try to explain that to the users :P

From the user perspective, it's not going to be a good experience if you rely on cold GNSS only on a mobile device (unless the user is already well aware of what to expect and how to behave).

Fair enough, I suppose. I think in the context of smartphones, continuously operating the GPS chip is simply not an option, but in other types of devices that isn't as big of a problem.

Fun fact: almanac and ephemeris data that makes (at least basic) AGNSS work is mostly available publicly, either from agencies that operate the constellations or from observatories worldwide. After reverse engineering the expected format I was able to feed the module I worked with with fresh data without having to use the proprietary (and unreliable) AGNSS service it came with.

Weird, I have a Pixel 8 and that's not been my experience at all. GPS locking still often takes a long time.

SatStat is a handy free tool for understanding what is going on "under the hood".

On the map page, the red "pin" and "circle" represent data originating from GPS. Blue color is data originating from Network Location provider(s). And you may also see "Fused Location", this is used by 3rd-parties like microG as one way to provide network location data to the OS.

https://f-droid.org/en/packages/com.vonglasow.michael.satsta...

> Raw GPS is universally unusable, requiring up to half an hour of waving your phone under clear sky,

Sigh. The increasing amount of unsubstantiated FUD finding its way on to HN recently is depressing.

Two glaringly obvious points to make at your nonsense:

1. What you describe is ancient history. It applied to old previous-generation devices with single-frequency chipsets (L1 GPS signals). However, effectively 100% of devices released within the last 3–5 years (and more in the case of certain manufacturers !) (a) have L1,L2,L5 GPS chipsets and (b) have multi-frequency GPS chipsets whch work not only with the "original" GPS but also Galileo, QZSS and others.

2. Even if we accept your argument, which we should not, then the rapid adoption of Apple's FindMy network (and the equivalent for Google) surely cannot have escaped your attention ? These serve to increase accuracy.

The reality is there is fast becoming no need for a third-party service such as the one Mozilla provided, because there is now first-class support for sufficiently robust high-accuracy positioning.

This is correct. I use GPS in RC planes and only the initial sync is that slow (when it has to download the ephemerides). Subsequent boots/syncs only take a few seconds, and you obviously get great resolution with just the GPS unit.

I'm assuming you're using a UBlox M8 or F9P or something similar? Most of the built units have a small battery or supercap that allows them to retain almanac and ephemeris data while powered off. If the ephemeris data is less than 4 hours old this will allow it to very quickly re-sync to the visible SVs and get a lock again.

It's quite impressive, though, at how the hardware has improved. For a complete cold start one of the challenging issues is to actually find a single satellite to listen to. For GPS there's 32 different spreading codes and each transmitted signal is affected by Doppler shift depending on its position and velocity relative to where you are on the Earth. You basically have to iterate through (spreading code, Doppler shift) pairs and run a correlator to see if there's a signal present. Contemporary hardware has a ton of correlators that can run in parallel which can significantly improve the TTFF (time to first fix). As soon as you find one SV and start to get a bit of almanac data you can quickly eliminate (spreading, Doppler) options from the problem space by dropping SVs that you know are not going to be visible.

Yep, exactly. And a mobile phone has an internet connection, which means that it always has an up-to-date ephemeris.

I use the Mozilla location services to have my Linux desktops and laptops automatically figure out where they are in order to properly apply blue-light filtering at night. How might one be able to integrate either the aforementioned Apple or Google services into the system to provide the appropriate location data? Or would it instead require the purchase of separate GPS units for these computers in order to have them know just where they are?

Dismissal of Mozilla being forced to bring down an essential service like this because the concerned post has some FUD in it, and that proprietary alternatives exist so what's the worry, is frankly disgusting. Especially when it's done via the blatantly immortal (but technically legal!) abuse of the legal system like this.

And it's not even FUD. The phone I use has no working GPS since it got degoogled, as it takes too long to get a poosition. Yes, it's older than 3 years, but maybe being able to integrate the mozilla location service (with something more adequate than LineageOS) was a factor in me thinking the ungoogled Android route was viable anyway. But here we are.

Not everyone wants to use the newest hardware.

I have a watch that I don't connect to my phone that gets GPS signal and lat lon and it works better than you describe. You made general claims despite your situation being specific to you - intentional or not, that's FUD.

I did nothing, I'm not parent.

Is that watch de-googled, running only with FOSS software, without a proprietary location service? How it the AGPS system mentioned in https://news.ycombinator.com/item?id=39725522 working, who provides the data? Was it stale?

If you can't answer that or if it's backed by (recently acquired) proprietary data, then you just do not know how GPS behaves in the scenario we are talking about.

> the rapid adoption of Apple's FindMy network (and the equivalent for Google) surely cannot have escaped your attention ? These serve to increase accuracy.

Do you mean Apple's and Google's Wi-Fi positioning databases? These are different from Find My: Find My is crowdsourced positioning of (themselves usually unconnected) trackers; the positioning databases are (at least partially) crowdsourced Wi-Fi (and maybe Bluetooth) signal to location mappings.

It's you who doesn't make much sense in this thread.

> released within the last 3–5 years

So, pretty much only the very fresh devices then.

> have L1,L2,L5 GPS chipsets

That's not going to reduce cold TTFF to seconds without online assistance (which, by the way, is not what MLS is about).

> work not only with the "original" GPS but also Galileo, QZSS and others

Same as above. None of the constellations or assistance data help when you can't see the sky.

> then the rapid adoption of Apple's FindMy network (and the equivalent for Google) surely cannot have escaped your attention

How is it relevant and how can it be used by anyone who isn't Apple or Google?

> The reality is there is fast becoming no need for a third-party service such as the one Mozilla provided

The reality is that I'm not aware of any currently existing alternatives to MLS that could be used by Geoclue or microG, which provide non-GNSS location on many phones out there, which is essential to provide anything resembling a good user experience when it comes to location. The opposite of what you said is true - there's a rapidly growing need for this kind of service as more FLOSS starts to be used on mobile phones.

> How is it relevant and how can it be used by anyone who isn't Apple or Google?

It is relevant because most people on the planet are not like you.

Most people on the planet use Apple, Google or indeed both.

Meanwhile, services such as the one Mozilla used to offer cost time and money to maintain. There's servers to maintain, software to maintain, CDN bills to pay.

If your user base is effectively a bunch of die-hard anti-Apple, anti-Google people, then its not worth the time or effort.

We are afterall, talking about location services here, not e.g. email where sure there's a market for non-Apple, non-Google.

For location services in 2024 there is little reason for people to not use what their device comes with, frankly.

Ah, okay. So you have no idea what you're talking about and don't care about what we're discussing here then, thanks for clarifying. There are devices that don't come from Apple or Google at all that rely on MLS and need a location service, regardless of whether you believe it's worth the effort or not.

AGPS provided by the mobile carriers is sufficient. It is downloading the ephemeris that is slow, not getting the fix. That isn't considered Location Services.

You could test this by disabling WiFi, downloading AGPS from mobile, and then driving to place without service or putting in airplane mode. GPS fix will be fast cause ephemeris is good for days and corrections are small.

Reply on news.ycombinator.com