Twirlip
Second Officer
Flight distance : 2461545 ft
United States
Offline
|
MARSAN Posted at 2017-7-16 00:21
I don't think that Android 5.1 was able to automatically put to sleep misbehaving apps.
Therefore Force Stopping the app manually appears to be the only solution, unless there is a possibility in the Developer options.
Regarding "back to square one with CrystalSky"-- No, not so much, I think it's much less of a concern there, for several reasons.
First, I don't live on the CrystalSky, the way I live on my phone. I use it only for flying the drone. Whereas my phone is how I stay in contact with the world and vice versa, and I carry it everywhere with the expectation that its charge will last all day. Phone running out of juice in the middle of the day when I'm far from home = serious inconvenience.
Second, unlike my phone which needs to be on all the time, I turn the CrystalSky off when I'm not using it to fly the drone. Basically, it doesn't matter if DJI GO 4 uses a lot of power in the background, because on the CrystalSky it isn't in the background-- it's the only reason I have the device in the first place.
Third, the CrystalSky has a ridiculously larger battery (one of the reasons for getting it in the first place)-- it hardly even notices the drain. I think the battery alone weighs at least three times what my phone does.
Finally: I'm reasonably sanguine about the future of software QA for running DJI GO 4 on the CrystalSky. Why? Because it's a flagship DJI product. This gets at how software production works (something I'm fairly familiar with, being a software engineer myself). No software is perfect; there's always *some* bug. How well software runs is a combination of how much money/effort a company's willing to invest in QA, and how hard a problem they're trying to solve. My impression of DJI thus far is that they're stunningly good at *hardware* engineering (the Mavic just makes my jaw drop), but when it comes to *software*... not so much. And I'm not super surprised at that. For one thing, trying to ship software that runs on a whole lot of platforms (e.g. iOS, plus umpteen flavors of Android) is a *huge* test matrix to work through. It's pretty clear that DJI is fairly Apple-centric, and their Android support is kind of an afterthought. It doesn't surprise me that their Apple software gets the lion's share of the QA love. (And I suspect that they write their software first for Apple and then try to port it to Android, which would stack the deck against Android, too.) I don't know that, of course, but it's consistent with what I observe (i.e. iOS software generally much more stable than Android) and it stands to reason.
But in the case of the CrystalSky: that's an actual, flagship DJI product. If DJI GO 4 won't run on the CrystalSky, then the CrystalSky itself is useless, and that gives their own hardware product a black eye, which means that presumably they'll be motivated to try to keep it working well. Furthermore, from a software engineering/QA standpoint, it's a much easier target than dealing with all the umpteen flavors of Android out there, because it's their platform: they control the hardware, they control which OS it's running, they control which version of DJI GO 4 is on it. Which means that when they need to test it, they don't have to test N zillion different potential combinations. It's a much, much simpler problem that way, which means if they only have N engineer hours to spend on it, they can focus those hours much more effectively.
What it boils down to is that I think CrystalSky will run the app better because, from DJI's point of view, 1. it's a much easier technical problem to solve, and 2. they can't afford for it not to run well.
And, in fact, a lot of that is why I got a CrystalSky in the first place. Not just because it has that nice bright display... but I was reading the forums for months before I got the Mavic, and I got the impression that their Android software is crap when running on phones, and my phone's an Android, and I didn't want to buy an awesome drone and have the experience ruined by subpar phone software. I made the bet that it would run better on CrystalSky than on Random Android Phone X, at least in the long term. Basically, I waited for CrystalSky availability before getting my Mavic.
Also, in retrospect, there's one more reason to prefer using a CrystalSky over my phone, which I didn't anticipate but in hindsight probably should have: privacy / security. The DJI GO 4 app on Android requires horrifyingly broad permissions. It's essentially asking for root-level access, it wants everything, even stuff that has nothing to do at all with the app itself (such as turning on my flashlight, or placing phone calls, or editing my contacts). It's obvious that DJI doesn't give a wet slap about my privacy per se, nor did their engineers take the time or effort to engineer the app to request minimal permissions. I don't see it as sinister, just lazy-- it's easier to just request global permissions. Which cares the willies out of me, because my phone is where I live. What I have and do on my phone, outside of DJI GO 4, is none of DJI's business. Whereas I don't mind whatever the app does on the CrystalSky, because it's their hardware, and there's nothing of mine there except the drone-flying stuff, which *is* DJI's business. So I'm a lot more comfortable from that aspect, too. |
|