Hacker News Reader

Comment by murderfs | Hacker News Reader

Yeah, pretty much all of the carriers cheat, and demand the ability to cheat, e.g.: https://android.googlesource.com/platform/frameworks/base/+/...

View on HN

>Yeah, pretty much all of the carriers cheat

Given that HN has an international audience,

1. Is this US only? I have never seen anything similar in EU or East Asia. I.e The Bar is an actual indication of the receiving signal, while it is no guarantee of speed which depends on contention, you will at least know you are in good signal area.

2. Is this Google Android only?

To answer #2, no. Apple played into ATT's lies when they were behind on 4G LTE deployments and labeled 3G HSPA+ as 4G. ATT is far more brazen than other US carriers in this scheme and should be held accountable. Apple also displayed this bogus logo on the iPhone at the time. Keep in mind, however, Apple had strong exclusivity with ATT during the early years of the product.

ATT seems to like to confuse the truth more often than other carriers as well. One big, recent, fiasco was supporting wideband audio (HD calling) and WiFi calling on Google Pixel handsets. ATT never said in print they didn't support it and made references to the fact that all feature were supported on their website. When you'd call them ATT support would claim it was fully supported, yet ATT did not. This was true for both OG Pixels and PIxel 2s for both at launch and well into those device lifecycles. I know because I had technical tickets opened for about a year with ATT before leaving them. The point being ATT likes to pretend a lot. Of all the carriers (have been customers of the major 3 for different products over the years) they are, in my opinion, the least trustworthy in that regard.

https://forums.att.com/t5/Device-Features/Pixel-Pixel-2-VoLT...

Call it what it is: fraud. They fraudulently deceive their customers for monetary gain.

Yeah. While European companies are by no means saints, there simply doesn’t seem to be this sort of constant screwing over paying customers around here.

Because there is certain line that shouldn't be cross in the old world. And this is called principle and value.

In the New world they are all for Marketing and Margins.

US doesn't have many protections for consumers. We are the country of companies and the populace doesn't have a real voice, just a fake one.

Agreed.

For US carriers its just business as usual. Technically the FTC would have to act on this but I think we all agree what side they are on. I really don't get why the US population doesn't demand better phone networks.

You want us to go into European, Indian, Chinese and African corporate scandals?

I have my popcorn out, do it! Just don’t start with VW.

To be fair, VW didn't try to screw over their customers per se; indeed the whole point was to increase fuel efficiency (direct benefit to customer) at the expense of increased NO2 emissions (a negative externality). Of course what they did was deplorable in any case.

Depends on whether you value money higher than life and health. Volkswagen is the biggest car brand in Europe - I'm sure they made health of their customers (as well as others) worse.

Not just in Europe, the Volkswagen group had the most sales globally in 2017.

How’s that Renault/Nissan thing going?

Don’t feed the trolls please.

I second bnt's idea. I cannot sleep and would love to see this play out.

We faked it in the phones I was involved in making. Android, EU.

My bars are for the signal, there is 4 bars but the 4G/LTE/3G etc...is right above it.

I'm on Android.

Ugh, I dared look at the code and stepped back in horror...https://android.googlesource.com/platform/frameworks/base/+/...

I've scrolled through that code, though I'm not a java dev, I don't really see much 'horror'. Can you explain?

Well, there is about 600 lines of boilerplate before you get to anything remotely interesting.

If the boilerplate is the horrific part, well that's just Java isn't it? The meaningful functions are straightforward and do what you expect. Describe the level of the first radio that returns non-zero starting with the most advanced technology. Or in the case of CDMA/EVDO, the lowest level.

Well no, Java isn’t the best of languages but my criticism stems from other issues, such as weak type system, certain inscrutably bad decisions about the runtime libraries and other, but this code is bad also according to any self respecting java developer.

An untyped (that is, it cops out to int) bunch of public static’s that should’ve been an enum. Daft combination of builders and constructors. Magic numbers. Adapter methods that should go into their own class. Validators that should also go into their own class (and even there, implemented differently, not as a parameterless method). Transformers that could be modeled as their own class (all these “in their own class” are for testabilty.) some of the names in these adapters suggest some other domain concerns (UI) are leaking into this model class.

IOW, it’s a mess by any standard

> An untyped (that is, it cops out to int) bunch of public static’s that should’ve been an enum.

Java platform code does not use enums since they are more taxing in terms of memory usage. Till Google I/O 2018, Google use to officially recommend that you should avoid enums in your app code if you can.

What's the motivation to lie? If service is claimed to be better than it is, then all that does is set expectations higher at times when service is weak thus increasing dissatisfaction and likelihood to jump ship.

Signal/network quality is what it is, signal bars are just there to set expectations for the user. Lying with them seems like a fool's errand.

I regularly hear friends/family/colleagues/strangers complain or switch service provider on the basis of “I get like no bars at home” or “I only get 1 bar reception in the office on <carrier name>”.

This stuff is relatively meaningless in a technical sense given even a full set of bars still tells you nothing about contention, backhaul etc, but it probably matters in marketing/customer perception terms for the carriers.

People only care about the bars after they notice calls dropping, and we assume the phone isn't lying. Then we get extra mad when we have 4 bars while calls drop.

As someone said, in practice what happens is that people like at the bars, not actual performance. Heck, I hear people who think the phone is messed up when their are full bars but no data working (because the bars are showing analog signal or other things). They rarely blame the carrier. So why should the carrier care?

What the fuck Android? Why are you helping mobile carriers lie to their customers?

Signal bars are meaningless anyway. Even a raw dB measurement wouldn't tell much about the service quality, so they might as well switch to a "coverage / no coverage" indicator.

To go a step further for GSM (I don't know precisely for 4G, but I suspect it is the same) the base station determines both handset dB/SNR/Quality and commands the handset power which usually is limiting performance. It is likely this setting that allows the handset to report its "number of bars"...

So if you don't let them lie on the handset... it's likely that the network providers would just start lying to the handsets at an even deeper level than they already do. Just get the marketing silliness out of the way at as high a level as possible so engineers can make our coverage better.

there are at least 4 values.

rsrp, rsrq, sinr, rssi

rsrq is probably the important one. You can clearly see usage of the cell tower with that.

Number of bars (1-2 vs. 4-5 or whatver max is) seems to give a pretty decent indication of latency and bandwidth to me. I get that it isn't perfect.

Seems to be a lot of situations where a bar or two is no service, and others where service is great. So as a determination of whether I can use the phone it is pretty meaningless.

Your experience will vary, but perhaps:

4-5 bars => 98% likelihood that everything is great.

1-2 bars => 20% quite good, 50% pretty decent, 20% quite bad, 7% bad, 3% no service.

How is something like that meaningless? That is immensely useful. You can also reason about it depending on the context. Downtown but inside a bar in a cellar? Or out in the country far from civilization?

That it isn't perfect doesn't make it meaningless.

In my experience, two bars perfectly indicate the threshold of ”there's service, but no service at all might've been better—at least there wouldn't be false hope.”

That actually helps explain an issue I had with my Galaxy Nexus. I worked in an area with spotty coverage, and it was extremely common that it would say I have full service coverage, only to have it drop to almost or no signal as soon as I used it to call. That issue was infuriating.

That can also happen if the phone switches between network types when calling - e.g. for a long time you couldn't call over LTE, so your phone switched to UMTS (3G) or GSM (2G) network to make and receive phone calls.

So you could have excellent service for data and still lose signal for calls.

That makes sense too. I thought that also at that time that LTE was relatively new, so in theory, the UMTS or GSM would have a stronger signal than LTE, which would have meant my calls would be fine.

I noticed this about phones, if you're using GPS and get a call and answer, guess what, no more GPS. So my guess is the phone cuts off the mobile data, and yeah then you'd see the real coverage.

Mobile networks are two-way systems. If you have perfect reception of the mobile network - let's say full bars! - then that doesn't mean the mobile network also has reception of your mobile device.

Most of the time, your phone is in RX-only mode to save battery. It listenes for incoming data every millisecond or two, but no more. But as soon as you actively use the network, things change. The basestation must receive your signals as well. This may suddenly and dramatically change the perception of "signal quality". Before it's RX quality only, then it's RX+TX quality.

With incoming or outgoing calls, your phone may switch to another technology. While it may be idling in 4G, large parts of the world rely on 2G for voice calls. As soon as the call starts, the phone switched to 2G and naturally shows 2G reception.

Interesting, yeah I don't know a whole lot about the underlying tech. I figure Google Maps uses my mobile data which explains why it sorta appears to have lost it's own darn mind.

CDMA phones (notably, 3G Sprint US, 3G Bell Canada, 3G Telus Canada) couldn't simultaneously use data (say, Google Maps) and voice.

GPS lat/long would work fine, but you won't be able to download turn-by-turn directions or map tiles/vectors from Google.

LTE (and 3G HSPA, such as AT&T, T-mo, 3G/4G/LTE Rogers Canada, LTE Bell Canada, LTE Telus Canada) don't have this behavior. Voice calls are packet switched (so you can do voice and data at the same time).

GPS is receive only, no mobile data is used there.

Assisted GPS might require data/mobile network connection for the brief period of time the phone figures out where it is, roughly.

Assisted GPS talks with the towers around, pings them for distance and gets their coordinates. It's happening on the GSM layer. It's also a no-data exchange IIRC.

No, I think it's only downloading almanach data for faster GPS.

It's one of the operational modes. On the other mode it calculates the location on behalf of the user and responds back with it [0].

In both modes data connection is required, so I'm mistaken.

[0]: https://en.wikipedia.org/wiki/Assisted_GPS

Google Maps uses mobile data to render the maps though do they not?

But how is this related to GPS?

Because that's what customers want.

Sure, they'll say they want accurate data, but at the end of the day they're paying the provider with the most bars and not doing even the most basic of independent testing, so clearly that's what customers want. In a free market, you measure what people want by what they choose to pay for.

No, the customers don't want faked information. You don't expect to people to independently test the contents of every item of food just in case the ingredients on the package on faked.

They simply don't know they're being lied to. In most other areas of commerce, that would be illegal.

A remarkably large number of US consumers definitely want fake information: as two brief examples, consider the way weather data in the US is reported (customers get angry if it rains on a “90 percent chance if clear skies” day and customers get angry if their car runs out if fuel when the guage hits “empty”

Interestingly, German consumers want to inow the exact level of petroleum spirits in a tank and will get irritated if they are lied to.

When I owned a German car, the speedometer read somewhat too high, compared to Japanese cars that were quite accurate. Also, on German cars I've driven, when they were stopped, the MPG gauge went to the maximum, rather than zero as would be correct. So I infer that German consumers want to be lied to about some things, just perhaps different things.

In EU, there's an unspoken rule for cars to report speeds inflated by 5%.

This is possibly based on an old research of SAAB which used a reverse logarithmic speedometer, giving feeling of exaggerated acceleration until 120-130 KM/h or so. Then the scale tightened so much that going significantly faster didn't move the speedometer that much. Also the needle hit top left quadrant much faster, hinting that you're going really fast.

Also, EU uses L/100KM concept instead of MPG. When the speed is 0, most cars switch to L/h scale. Toyotas also show an infinity symbol when in MPG mode while the speed is 0.

For anyone interested below are some examples:

1- https://upload.wikimedia.org/wikipedia/commons/5/5c/Instrume...

2- https://i.ytimg.com/vi/V0RyHPM0mBI/maxresdefault.jpg

I remember the numbers right, but it wasn't reverse logarithmic scale. It was a hard transition at 140KM/h.

Are you sure they displayed mpg and not l/100km?

Dashboards are user or factory configurable. Our Ford Focus had a button to change between L/100KM and MPG.

My friend's VW Bora's dashboard was OBDII configurable. The options were Japan/US/EU IIRC. Setting to US not only changed it to MPG scale, enabled the door-open beeps. :D

Well, I wouldn't be surprised if the software internally always computed l/100km and the converted to mpg for display.

It's much more direct for car to measure consumption as L/100KM. However the conversion is done by the instrument cluster on the VW. We were connecting to the instrument cluster via CANBUS and changing the setting directly on it. Also the beeps were handled at the cluster, not at the central config.

Car technology is weird...

The car manufacturers always make their speedometers show a little bit too much (like 10%) to make sure they never show too little (or they would get sued badly and get their cars taken of the streets). We _accept_ that the car companies do this for legal/security reasons, which is very different from _wanting_ them to do it. I would be angry if they did the same if the fuel indicator.

No they don’t. I have a 2005 ford with a speedometer that matches GPS all of the way up to 85 mph (haven’t tested higher).

I had a 1996 Ford Mondeo that showed 10% more. Then I had 2004 Ford Mondeo and it was 5% more. I sold that one and bought a 2008 Ford Mondeo and it is 5% more. My other car - 1998 Ford Fiesta - is 10% more. I also drove a 2011 Ford Mondeo and it showed more but I hadn't had it long enough to care to calculate how much exactly.

I tested all the cars except the Fiesta up to 230 km/h and the 2008 one up to 260 km/h (it's the 2.5T).

All the cars are/were German except for the Fiesta, that one is Czech.

Maybe it’s something they do in Europe but not the US.

Maybe your car GPS is reporting higher? All cars I've driven were reporting %5-%10 more.

It’s not a car gps. It’s a receiver on a phone. That can’t be wrong by that much or the position would be wrong.

Japanese customers, too. Back in the day, there was a gentleman's agreement among Japanese automakers to limit power output of JDM cars to 276 hp.

In reality, Japanese performance cars were all spectacularly underrated. Your "276 hp" sports coupe is probably making 350 hp. It was more of a gentleman's agreement to lie about power figures than anything else.

Customers can and do independently test their products when they care, from the UL to review sites to kosher certification to gluten-free certification to animal welfare standards at grocery store deli counters. The free market is adequately capable of answering these questions.

When customers actually care enough to pass a law, they can. That's why certain practices are illegal. Other ones aren't because they don't want their taxes increased to fund investigations they don't care about. When the government starts enforcing protections for people's own good because they know better, you no longer have a free market.

Technically, the government determines what qualifies "certified gluten-free", at least in the USA[1]. It probably belongs more in the "when customers care enough to pass a law" section than the "self-regulation" section.

1: https://en.wikipedia.org/wiki/Food_Allergen_Labeling_and_Con...

They are getting faked info. You run out there and tell them, see how many switch.

You're confusing what customers want with what satisfices them. If customers saw one bar and 3G all the time, that would lead to more complaints and worse image of their service. The carrier is doing what's expedient to reduce support calls and placebo their way into "better" service at the least short-term possible cost: lying to the customer.

The trick is to trust but put people in prison when you find out they betrayed your trust.

Because the alternative is for every single OEM to reimplement this on their own, probably in a worse way. Everyone (including Apple, who have far more leverage over carriers) is basically forced to comply with the carriers here, so there's really no other option.

Playing devil's advocate: it's purpose could be to help them not lie. Imagine if there are circumstances where you have good signal strength but still poor quality; a carrier could adjust the bars displayed to be more realistic.

At least you admit you're advocating for the Devil!

We could hope a cell phone company would be so honest as to make sure their signal indicators show possibly lower values.

They could, sure. They could also give you a free pony with every phone.

Because mobile carriers and other large businesses have always been their customers, not consumers. Google’s only incentive for helping consumers is that it might give them more data for their large business partners.

What's your explanation for why iOS does the same thing?

Apple as a company doesn’t care about the truth and never has pretended to (which does suck).

This echoes of the "Antennagate" problems where apparently Apple's algorithm miscalculated signal, there was a big bruhaha and Papermaster was thrown under the bus.

$$$$$$$$

Really?

https://www.reuters.com/article/us-apple/apple-says-iphones-...

It wasn't Android. Apple did this deliberately, and only apologized when they got caught.

It's a ritual to bash Android and Google on HN, and Apple gets a pass every time.

Apple also released a patch which showed "4G" symbol in status bar for AT&T when on HSxPA signal. In europe and everywhere else, the phone kept showing 3G as other phones.

It’s a per-SIM carrier configuration, when you’re abroad with a US SIM you’ll also see 4G/LTE.

Ahh, good to know.

Does anyone know if this cheating is permitted on iOS?

This seems like Apple was borderline admitting the bars are a joke:

https://www.wired.com/2010/07/apple-claims-iphone-signal-los...

Reply on news.ycombinator.com