Weekly Update from Alex - 05/19/16


(Alex) #1

There have been some solid improvements this week that I am excited to share with you.

Platform Updates
We pushed a platform update this week with a few key changes I would like to highlight. First, in regards to device limits in the mobile app, we heard you, and I am happy to announce we increased those limits from 150 to 300. We also made some changes focused on developers and making the development experience better. For more information on these changes and more, follow the discussion on our Release Notes thread.

Rooms Performance
We added some caching to our rooms service to reduce load on the platform. This update was pushed out yesterday and we have been monitoring the metrics on throughput to the database after this change. We are seeing reduced load on the section of the database where Rooms live. The graph below shows the more stable and lower throughput to the database since we improved rooms caching.

Stay tuned for more updates next week as we continue to dig into the platform and improve it everywhere we can. Please let us know how your daily usage has been. We watch and use these threads to prioritize the following week’s goals.

Finally, we’re pumped to be putting on our first hackathon in conjunction with the Samsung Accelerator group. We’ll be hacking all weekend with hundreds of people to create awesome experiences. We’ll be tweeting all weekend long to show off everyone's hacks and if you are in the area we’d enjoy seeing you there! More information on the hackathon can be found here.

-Alex


(Bobby) #3

When all the threads that have been updated in the past few hours are basic set up questions, new app releases or outgoing developments, that's a good indication that your team's efforts are paying off! I personally haven't had any issues in over a month now. Not a routine being missed not a schedule being skipped. I must admit, that is a personal record in almost 2 years of using SmartThings. Thank you!


(Geko) #4

Platform stability has definitely improved. Hope the mobile app will be the next focus of the executive spotlight. It really is a sore spot, particularly on iPad.


(Alex) #5

I hear you. We won't relent in this focus on the basics ever again. So, yes the mobile apps will be getting lots of attention this year.


(Marc) #6

Thanks @alex. Would this include fixing mobile presense reliability issues? Many of us on the forums have had to resort to key fobs, Life360, IFTTT, WW-DRT scripts, etc. just to work around this weakness.


(Bobby) #7

I know the solution to that, but how do we convince @alex to turn on the Bluetooth for us?


(Marc) #8

Won't help us v1 users :slight_smile: Also, Bluetooth is very limited in range, and while it can serve as a backup, it won't help with garage door and lock automation for a lot of folks.


(Ash (www.smart-dots.com) / Ashutosh Jaiswal) #9

The reliability has certainly improved drastically! THANK YOU VERY MUCH for listening to all of us.. and prioritizing things that matter the most...We are all very excited!!! Thanks!!


#10

Hopefully including a real fix for voiceover navigation. :wink:

As far as focusing on the basics, I regard the voiceover failures as a "canary in the coal mine" for the mobile app as a whole in two important regards:

One) Testing Methodology

If the mobile app is broken 100% of the time for a basic feature in a way that can be demonstrated with a screenshot and there is an official announcement that that specific problem has been fixed when in fact the same screenshot shows that it is broken in the same way, there's a problem with testing procedures. Even with the concept of testing itself.

From the Canary standpoint, it looks like you are testing the fix rather than testing the feature. That is, you are testing from the programmer's viewpoint rather than the customer's. It's like checking to see if a word is spelled correctly without checking to see if the sentence makes sense.

Two) Discoverability

One of my old professors used to say "The hardest thing to remember is what you didn't used to know." Meaning it's really hard to look at a page in a book and see it the way it looked to you when you weren't quite sure which letter was the "b" and which letter was the "d." We see this time and again with the mobile app.

It has really powerful features, but it's almost impossible to find them unless you already know that they're there and you already know what they do.

Two obvious examples which people ask about all the time in the forums are grouping devices so that you can schedule several lights together, and the classic "I just got home" scenario where the porch lights turn on for five minutes and then turn off again automatically.

Can you do both of these with SmartThings? Absolutely.

If you don't already know how to do them, can you discover that functionality easily? Not likely.

The Canary aspect comes in here because trying to find things with broken navigation becomes that much harder when the navigation itself obscures what is there.

I understand if fixing accessibility is not that high of a priority. However, with regard to a focus on the basics, I would respectfully suggest that defining discoverability as a basic would significantly improve the usability of the platform for many markets.

JMO :sunglasses:


#11

I don't understand why people think that's Smartthings issue to fix. They probably read the current location from what the phone gives them. If it is inaccurate there is nothing they could do other than keep your GPS always on and your device awake and drain the battery.


(Michael Hess) #12

GPS is very low drain compared to cell or wifi radios. I've not experienced this issue personally as Life360 works flawlessly for me, but having a background process for the app checking every 5-10 seconds seems like it should suffice. I have no idea what or how it does it right now.


(Bobby) #13

@michaelahess can you chime in to this thread when you get a chance to share your thoughts on why you have good experience with L360 while others don't ?


(Clayton) #14

This is not at all true. The general rule of thumb is that the further away whatever the radio is connecting to, the more power it will take. Therefore, Bluetooth would drain the least, then Wi-Fi, then cell, then GPS would drain the most. This may not always be the case but it is a good guideline.

That being said, your phone doesn't always use actual GPS for location. GPS provides the best accuracy, but often times a relatively close location can be obtained from triangulating from cell towers, or a very general location can sometimes be obtained through a Wi-Fi connection.

I guarantee if your phone was constantly connected to 3+ satellites using the actual GPS antenna your battery would be dead in a few hours if not sooner.


(Michael Hess) #15

GPS doesn't have a signal strength issue like your other radio's, it just receives "pings" from the satellites, it's not active like BT or WiFi. I've done very long (weeks) tests with battery monitoring software on a multitude of phones with GPS on and Off, same with BT. GPS makes no perceivable impact.

WiFi, BT, Cell, all those radios will ramp up power use for TRANSMIT when you get further away. GPS does NOT transmit.


(Jason "The Enabler" as deemed so by @Smart) #16

Well, I'm definitely impressed now.

Motion sensors are active. Rules show true. Lights show on in the app.

Guess what's occurring in the real world? ? ?
That's right....NOTHING!

EDIT
no control over anything. Alexa, says ok. Nothing. The app says it is on.... Nothing.

Super!


(Clayton) #17

GPS absolutely has signal strength issues. Try connecting to satellites from an underground parking garage, or downtown in a big city with lots of tall buildings. You'll be lucky if you can connect at all. Turning GPS on/off on your phone is NOT a valid test. Just because GPS is turned on, does not mean that any apps are requesting your location at a high enough accuracy level for the GPS to actually be connecting to satellites. The notion in question,

is completely valid, because if the app were constantly requesting your precise location using GPS your battery would drain extremely fast.

This is just the first result of a quick Google search corroborating this: https://www.quora.com/Why-does-GPS-use-so-much-more-battery-than-any-other-antenna-or-sensor-in-a-smartphone

I'm not going to argue with you about this, I don't care about changing your mind. But I don't want other people taking this bad information as fact. For anyone who doesn't believe me: Start a GPS navigation app on your phone and have it start a route to any location. This should be the easiest way to force your phone to be constantly updating your precise location using actual GPS. Then leave it running and see how quickly your battery drains. If the app will run in the background you can even turn the screen on your phone off to be sure that the battery drain is due to GPS and not the screen being on (which is actually a bigger battery drain than any of your radios.)


(Kevin) #18

Yup, I think theres an outage. Noticed it about 15mins ago when no lights came on with motion, nor door sensor. But how are we to know...cause status page says A OK. Devices that I've had 100% reliability arent controlling right now... AWS status looks good other than some api reports... so who knows


SmartApp triggered with virtual switch not working from mobile device
(Michael Hess) #19

Your link proves it's not the GPS, hence my point about a well programmed app using GPS is not a battery hog.

"This is again due not to powering the GPS itself, but by preventing the phone from going to sleep."

I use a GPS logger for running/biking, it takes a positional measurement every 5 seconds, there is no noticeable drain because it properly puts the phone back to sleep between readings. Anything keeping the phone awake continuously will kill your battery, not just GPS.


(Kevin) #20

And my control has returned. At least 30mins if not longer. Its the middle of the day so not like i have lights coming on and off. But yea, i was 'down' in terms of control thats for sure.


(Clayton) #21

Right, and the point of the comment that brought this all up is that to increase accuracy you have to increase polling which would prevent the phone from sleeping as long.

Not to be pedantic, but I don't think apps can put the phone back to sleep, I think they stop preventing it from sleeping, and it's up to the OS to sleep when it can. This would obviously be platform dependent but I'm pretty sure iOS and Android work this way.

I completely agree, and this is by far the biggest drain of battery on any platform.