I don't understand the SmartThings/SmartApps ecosystem -- thoughts and feedback after several weeks

Just tested with my Windows phone… Same ~1 second update time in the SmartThings App.

Do you have a special VIP support email address to contact? My experience with support has been anything but “outstanding”.

I would say one of every five issues I report to support results in a reply that is worth my time and actually informs me. Usually I find myself having to train support staff on how SmartThings works.

There seems to be a significant gap and time lag between what gets reported to support and what the engineers get notified about and fix. I’m pretty sure this is part of the recent problems. There’s a lot of lag of many days in issues getting attention. And I get replies that indicate support staff are supporting PR and marketing events and therefore don’t have time to monitor and handle issues that would have prevented or at least mitigated the platform outages and issues of any given day. It’s part of growing pains of an alpha stage product. You can’t prioritize PR events over a stable product IMHO.

Thanks friends for the link in response to my comment about support issues. They are working on it, which is good to see!

1 Like

Hey @dlee,

Just providing some insight and perspective inside the company:

We are so avoiding that as much as possible, moving forward. At a point, when you have limited resources, there’s only so much we can do. Do we take the engineer who’s stabilizing the platform because of these obligations we have to do, or the support person, who also has a great understanding of the platform. We certainly aren’t sending them to GDC to party at an experiential event; these events are generally industry events that we have to attend.

When you had a team of ~7 support staff, the time/experience does get delayed when we do sacrifice a support staff member for these obligations. We are continuing to grow that portion of the team, (yay! There’s about 3+ new people on the support team, which brings the count to 10) and getting them up to speed as quickly as possible.

Also, just a clarification: Support are not the same people who would can prevent or at least mitigate the platform outages and issues on a given day. Support is customer service, customer facing, and well, Support. Our engineers are the ones involved with that.

Regardless, we’re hiring on all fronts, so PR/Events/Marketing efforts doesn’t affect the rest of the world, because I agree; a stable platform is much more desirable.

:slight_smile: They leave PR events for people like me, who is not involved in support or engineering. Although, I admit, engineers and support staff members sometimes show up to our local events : However, on their “off time”.

3 Likes

Yes, yes, a thousand times yes! I’m a CTO of a high-performance software company and I echo everything rogersmj says. The mishmash that is the SmartThings app is completely unintuitive and incredibly frustrating in all the ways he discusses, and more. It’s enough to drive me away from the platform altogether (though I admit, I am still here).

Is the API open enough that a third party developer could put a better app together, based on such a rule-builder and simple basic ideas like triggers, events and actions? Maybe even web-based rather than phone-based so you can set up your smart home on a decent size screen?

(Oh, and BTW: the current IFTTT channel doesn’t support Minimote or other button devices, so for me it’s 100% useless. Oh well.)

2 Likes

It can work fine with the minimote. but you have to use a virtual switch for a “man in the middle” protocol.

So the minimote button is assigned to turning on a virtual switch, that virtual switch coming on becomes the “if” to IFTTT and you’re off.

Because cut-and-paste is very difficult for me (I’m quadriparetic and depend on voice recognition software), I use IFTTT a lot, probably more than most smartthings customers. Virtual switches are the key.

As far as the UI, yeah, it sucks. And I have no idea why it isn’t better. I’ve been waiting a year. I really like the SmartThings vision, I think the staff are great and the community is fantastic. But a lot of it still just feels like proof of concept to me. Still, my SmartThings makes my echo hundred times better, and I love my echo, so I can’t say I’m unhappy. Just frustrated.

There are two topics that might help. (These are clickable links.) The first one is for people who don’t code at all, and highlights some of the third-party solutions that community members have developed to fill some of the UI gaps. It includes a dashboard solution and two different rules engine apps, one for iOS that is sold through the App Store and a free one that is browser-based. All developed by independent community members on their own time. So obviously it can be done. :sunglasses: All three of these are very popular.

The second topic is more for “power users” and details some of the many different scheduler options that are available in SmartThings even if there’s nothing official to tell you that they’re there.

I realize that’s not an answer for why the platform doesn’t have an absolutely killer 21st-century mobile app (which could still communicate with the hub even if the cloud connection is down, something that would make a big difference to me personally), plus a web-based rules engine. But maybe it will still be of interest.

3 Likes

@Gary_Oberbrunner I was starting to reply but @JDRoberts beat me to it. The entire philosophy behind this platform is that it is open so it will be a mishmash of things for a long time because so many people contribute and because it is so open.

If you want something that is more streamlined you’ll need to go with something with a little less flexible but is more reliable.

IFTTT does work using Virtual devices w/ MiniMote, like @JDRoberts says.

I built http://www.simplerulebuilder.com/ because I felt the same way. Simple Rule Builder is a web based rule engine that allows you to define rules based on a Trigger, Condition(s), and Action(s). It’s free to use, feel free to check it out.

2 Likes

I appreciate that, but IMHO this is a beautiful description of the problem. It’s one thing to be “built for developers”, heck I am one. I know what that means. It’s another thing to be “built by accretion” which appears to be the case here. Every little easy thing appears to be complicated and there are way too many almost-the-same-but-not-quite ways to do something. So the chain proposed here is minimote → ST virtual switch (edit/adjust on phone) → enable on IFTTT (web) → talks back to ST → does action. Really, that’s just not maintainable. If I add another light in 6 months or want to switch to button 3 instead of 1, I’m not going to be able to remember how I did that. Sure I can make myself a log of everything I did, but really? It doesn’t have to be that way. OTOH, I just got a pointer to https://www.simplerulebuilder.com/ – this looks like exactly what’s needed! (Haven’t played with it enough to know yet.) ST, can you just buy the tech from @JoeC, the simplerulebuilder guy? :smile:

BTW, I used to be an amateur theater lighting designer - they’ve figured this stuff out. Complex scenes, fades, event triggers, design-by-example, all that stuff. It doesn’t have to be hard.

3 Likes

Thanks! This looks brilliant, @JoeC. I bet I can delete most of my built-in stuff (if I can find it all) and replace with this.

No, I’m not looking for something more streamlined in the sense of less flexible or more hidden, I’m looking for a UI that makes sense, gives me more control, more transparency, and more visibility into the automation I’m constructing. ST’s current UI is only flexible in a sort of Stockholm-syndrome way: see how excited people get about “virtual switches”. They’re a hack workaround needed due to the current UI architecture. Nothing in the hardware or device layer necessitates them. But hey, maybe @JoeC’s simplerulebuilder will be the beginning of the new and shinier world.

2 Likes

@Gary_Oberbrunner I understand nothing in the hardware makes the virtual devices required, but making complex setups with infinite combination of devices and use cases it is difficult to make the software “simple” for that and not reduce functionality.

It is similar to iOS to Android, iOS is much simpler to manage and maintain but you don’t the flexibility you do with Android.

JoeC ui looks great, and it looks very similar to SmartRules http://smartrulesapp.com/ as well. But when people something more simple… you want more control, more transparency and more visibility into the automation, just simple build your own SmartApp. That is the power of SmartThings giving you that flexibly and then you can have told control over it.

People want both control and simple, and that is very difficult to accomplish to balance those 2 competing interests.

2 Likes

Well… that’s debatable…

“Any sufficiently advanced technology is indistinguishable from magic…” (or whatever the exact quote is :wink:).

SmartThings launched with a different “philosophy” in regards to automation than every other Home Automation product before it, both highly proprietary ones and very open ones. Perhaps it was (is?) an attempt to differentiate, be more open, and be the “first” Smart Home platform rather than “Home Automation”.

These products successfully providing a lot of useful functionality using rule-builder interfaces.

And, in fact, the very existence of SmartRules App and Simple Rule Builder (and SmartTiles for a simplified dashboard view) proves that the Platform can have both complexity and simplicity. The User Experience designers at SmartThings have just failed (or chosen…) not to deliver this magic.

3 Likes

I agree with most of what you’re saying, but not the virtual switches comment.

I have looked at many other platforms. And considered a lot of them. The main thing that has kept me with SmartThings is the ability to use virtual switches to make connections between devices controlled by smartthings and other services without having to buy actual physical devices.

Insteon has a much better UI than SmartThings. It has exactly the same echo integration as SmartThings does with one exception: The ability to use virtual switches with smartthings.

Because of those virtual switches, I can now use echo for granular control of my home theater. I can turn the volume up or down by voice, I can pause Or rewind by voice, I can switch to a channel favorite like ESPN TV. All because of the use of a virtual switch to stand in for a routine.

Even Amazon integration for fire TV can’t give me the same granularity. I can’t pause inside of Netflix using fire TV the way I can using smartthings, a virtual switch, and harmony.

( i’m quadriparetic, so I really value totally hands-free voice control. I can’t turn my television off without it.)

So I think this is one of those “don’t throw the baby out with the bathwater” situations.

I agree that virtual switches are often used as hacks to jump over gaps in the SmartThings UI.

Separately, they are used to create interfaces to external systems, nothing to do with the UI at all. And that’s a very powerful and useful tool. :sunglasses:

Submitted with respect.

3 Likes

All the more reason that SmartThings should make these an integral part of the User Experience… i.e., a dashboard page (or some logical section of the mobile App) where these switches can be created on the fly, organized and, using the Platform’s inherent knowledge of other binary states, programmed to trigger other actions. Yes – there are some SmartApps leaning in this direction, but nothing quite beats official integration especially for slightly complex concept. SmartThings can do things internally that still are cleaner than SmartApps and then would be fully supported as well, such as synchronizing an open/close action with a on/off Virtual Switch, etc…

2 Likes

Ah, OK – just goes to show I have some things to learn. I guess it’s the fact that I have a controller, with, you know, switches, and adding a virtual switch to a switch just seems like needless indirection. But I totally see how they could be useful in a different context. Thanks!

3 Likes

Wow, this thread is food for thought. It makes me think back to 25 years ago when I was exposed to things like Smalltalk, LISP and Prolog and spent time building discrete event simulation models and Expert Systems.

I am slightly surprised that these issues were not apparent to SmartThings from the start. It seems they are learning some lessons already learnt by others. I think it is vital to have a proper model of how HA can be worked in a way which supports simple things but is extensible. It surprises me that SmartThings did not think all this through with thousands of man hours at the whiteboards before they started building a platform.

I think they desperately need a more sophisticated App that uses a large screen real estate and people can run locally e.g. On a NAS. I am amazed there is no App except for mobile phones. You can’t even see more than 6 things or 6 SmartApps on the screen at one time for goodness sake. You don’t seem to be able to sort them or put them into groups either.

Long lags going out to the cloud and back is not an attractive feature.

Don’t want to be too negative as I am actually quite enthused at the moment with the potential of an open platform and increasing availability and decreasing cost of zigbee and zwave devices.

1 Like

Who would have paid for that?

SmartThings launched as a pretty darn successful Kickstarter Project, but ( $1.2 million - product costs) / “thousands of man hours” = far too little salary for anyone.

Have you heard of MVP (Minimum Viable Product)?

1 Like

Normal round-trip to cloud is a couple of hundred milliseconds, at most. There are other sources of lags in the system, most notably from z-wave devices that don’t report state changes promptly.

The complexity of ST is well beyond anything going on 25 years ago. You’re looking at a vastly distributed event driven asynchronous realtime system that hangs together pretty well for the most part, considering its complexity. Somebody did something right here, more than may be obvious at first looks.

3 Likes

Fair points both, I take them onboard and was undoubtedly being over simplistic and provocative.

As some defence…
I guess I’d have thought $1.2M would translate into 30,000 man hours.
Why did I see comments that Smartthings are moving some processing onto the hub if the cloud is fast enough ?

1 Like

OK. I suppose its a pity the Kick Starter backers didnt get any of the $200M Samsung paid.