In case y'all missed it, it's Engadget reporting more specifically on the problem. Including what was disabled by default and what the real problem is.
Aero, they spoke English in that article, but since you speak Greek, I assume you know what you're talking about. Do you think this Engadget article is accurate? Is a software fix even possible?
I wrote directly to the editor after the samsung "clarification". I bet many people did. At least they are now breaking it down more.
I think the first issue is still more than fuzzy:
Samsung tells us that it's a new Google mandate that Android devices be shipped with the "use wireless networks" option disabled, which means you're relying on traditional GPS alone to determine your location.....get our location with 1,000 to 1,500-meter accuracy
This seems to me that with that issue they are saying they are only referring to Provider location, or triangulation from land-based cellular towers that gives you very rough fixes (about 1000m to 100m) being default off. (why?: data charges on some networks and battery life)
But they are not at all discussing any affect (if any) this may have on aGPS,which is about something different: giving the GPS satellite initial acquisition a boost by having the network tell the smart phone where to look for the satellites, and some other information about the satellites that should be in view, and then having the device calculate position from the satellites alone (aGPS).
on the second problem, this, if accurate, is in the mid range of troubling.
All Galaxy S models seem to be having trouble turning GPS reception into coordinates, even when the phone is able to see four or more satellites in view (four is the minimum you normally need for a precise, three-dimensional lock).
On the very good side the implication is that it is not a SNR (signal to noise) problem. That is good since a hardware issue affecting reception of satellite data per se would be nearly irreparable..
On the bad said the implication is that is also not cache or agps issue either which would be the easiest fix.
Also on the bad side is that the answer makes little sense*.
I don't want to speculate too much because only some of the statement maybe sammy, and/or may have passed though a PR/legal person, and some maybe extrapolation from engadget.
I think the second problem is more likely some kind of conflict in the way Sammy is prioritizing what localization (network or agps/gps)the device will be attempting. Since a coarse estimate of position is avaible from tower triangulation, their api, or adaption of an android api or code maybe written in such as way Galaxy doesn't try the actual GPS often, ore insistently enough. In other words some accuracy value set to low may be being is being satisfied (or waiting to be satisfied not triggering gps at all) by triangulation when it is not really accurate as the phoens actual GPS can do. [Why would such a accuracy value be set to low? Battery life for one. What you want google maps to show when you are looking for local information (banks, theaters etc) is not as demanding as what you wan t a driving navigation program to utilize. But true GPS may draw 500ma or more and you can see that would mean -2 hours batter life.]
The way android is supposed to work is that most LBA's (applications that need location) are written to access GPS as a preference, and then provider location as a second choice if no gps fix comes within a certain short a certain timeout. It looks like this is being played with for battery reasons. Even with null GPS value initially , LBAs are supposed to then poll the device for actual GPS fix at an interval regardless (hence you might start with an appropriate location and then over a short period of time get a real fix).
overall it is disconcerting that the problem exists, but at least engadget is getting a response from Sammy and sammy is clarifying that they are working on a fix for item #2 is actually quite heartening..
*It seems unlikely in the extreme that very simple, ubiquitous tasks of calculating position is any problem at all since literally hundreds of millions of devices do it, the mechanism are decades old, and all the methods are simple and wold not be affected by aGPS issues, driver conflicts, or even software bugs. the simplest part of building any GPS receiver is calculating position once sats have been acquired.
simplified explanation location mechanism that can deliver approximate or accurate fixes to application:
- Triangulation/"Provider location": This is on all phones, dumb and smart. this uses only the handset and carrier towers (and computation of location is usually done by computer at the carrier). It give fixes from several kilometers to 100 meters typically. On dumbphones this exists so emergency 911 services can locate the phone, and the actual location may not be avalable to the user of the phone. On smartphones the rough estimate can be useful to the user and applications since it a) works indoors or b) can help in situations where less than 4 sats are visible to the device and c) usually consumes way less battery than the GPS.
- Standalone GPS: the system most portable and removable car GPS use. This involves is the device and the satellites in orbit only. Most smartphones have this and aGPS
- aGPS: here the carrier sends data to the smartphone that help the smartphone locate the satellites faster. The smartphone will then use the satellites to calculate its position.
The best smartphone have all three (and good drivers and APIs to select which one in what case). My treo Pro does.
Triangulation as discussed is mandated by law, and is good where gps is not available or for extended periods without external power.
Standalone GPS is often slower initial position fix (TTFF) than aGPS, but standalone works on 95% of the globe that is outdoors,meaning say overseas or in extreme rural areas, which is why the best smartphones try aGPS first and fall back to standalone.
Facebook
Twitter
Instagram