Announcement | Changes to our Legacy SmartThings Platform

We have some samples over on glitch

3 Likes

And if you want a tutorial for running one, here you go

3 Likes

I listened to the podcast and here are my takeaways:

  1. Smartthings/Samsung has big plans for the future of this platform - however for whatever reason they seem to not be sharing very much information besides the Phased approach above.
  2. By getting rid of groovy and forcing you to use their own in house API – this will make it harder for newcomers to just jump right in and develop a functioning app and let them attract new application developers (eg big $$) to develop a brand that is “Developed by SmartThings”…
  3. They don’t seem to initially want to tick off the development/user community, (since there are 62 million active monthly users) however I am not getting a lot of warm and fuzzies from their transparency.

I believe the reason they are doing this is because they are not making any reoccurring revenue from their current model besides the sales of the initial hub. Because unlike Google (a search company) or Amazon (a shopping platform) or Apple (a hardware company) they have no way to monetize the costs of their infrastructure – in its current form I do not think Smartthings can be a revenue generating product line can it? I mean back end servers cost $…

Thoughts?

2 Likes

sure they do: sell more TVs and appliances. SmartTHings is basically an R&D division of Samsung that works to sell more of their phones, TVs, Fridges, Washing machines, etc.

4 Likes

Yes but does this make it harder for geeks who are just hackers and not actual coders to develop in the platform? I know a little python and a little java – however I could pretty easily have developed an app (even though I didn’t) because groovy was easy to understand…?

Yea but when you buy a fridge – do you say… i want it to work with smartthings? – I’m not sure…maybe us geeks do – but I am more likely to be like does it actually work right? Or is it a quality product… I see your point though

1 Like

I will have to say… If I buy a fridge or dishwasher or dryer I am more likely to ask if it works with my echo show… I think everyone is creating silos here?

1 Like

I mean a raspberry pi is capable of running different coding languages so that’s not really an issue. The “SmartApp manager” really would just be a front end for pulling down developer’s code and installing and running. Honestly, if smartapps were just docker containers it would probably be pretty simple and straightforward. Then people have the choice of local hosting, self cloud hosting, or developer hosted (if offered).

Developers aren’t going to want to developer if its hard to get users to install and run their projects. Having a simple, straightforward solution would show support for the developers, push additional development of features from the community, and probably draw more customers in. Today you can setup github integration in the IDE and its very easy to walk someone who doesn’t know how to code through the process of doing that to pull in device handlers or smart apps.

Right now with this new model it seems developers will have to pay for hosting their smart apps, which will mean they will have to start charging for it which will in turn cause less people to want to use them. This will then cause the developers to work on projects less / at all on the platform.

Maybe this is only for a small percentage of the user base, but I know the reason I recommend SmartThings to others when they ask for starting out is because of the flexibility and actual capability added by the community. Even to those who are just getting started and may not be the most tech savvy.

I’m perfectly fine with ST no longer hosting things. It makes sense, and it’s a cost to them but they really should provide the tools to allow the community to easily transition, continue to use, and help spread its growth to others.

3 Likes

This is exactly why it will turn people off… too complicated and too many pieces to try and get all working… It might be good for Samsung and it probably is way more powerful at the end of the whole transition process… but I think it will be hard from people to get started that ARENT developers

1 Like

Agreed 1000% – same here too I am afriad we might be loosing that

1 Like

Remember that everyone who buys a Samsung smart television or a Samsung smart appliance is using the same platform and the same app. Those 62 million active monthly users are the ones using the app. Not the ones using SmartThings automations. And a much smaller percentage of that small percentage are the ones using custom code.

So the “62 million users” are not the same as the “development/user community.“

This forum, for example, has less than 85,000 registered users. And it’s not like there are millions of other developers out there who are not members of this forum.

6 Likes

I’m not really much of a coder myself, as anyone who has read my code can probably tell. To bend a phrase from the classic “Real Programmers Don’t Use Pascal”, I can write Fortran programs in any language.

I don’t know about AWS Lambdas, and I think whoever put them up front and centre in the ‘new’ documentation a year or two ago needs to have a word with themselves, but I think it has been realised that they might frighten the horses and that WebHook Endpoints are probably a better starting point for the less experienced programmers. To borrow another phrase from the article I mentioned previously, I personally feel that some of the examples of using the TypeScript/JavaScript SDK that I’ve seen “look more like transmission line noise than readable text” and I do worry that that could put off those who are more comfortable with more verbose procedural programming. On the whole though, whatever the language you want to use, I think users will find it a whole lot easier.

2 Likes

One at a time, yes. Not all at once. You’re going to run out of resources pretty quick.

So, yes, it’s still an issue.

1 Like

I would spin it differently. I would say that developers will be able to monetise their efforts by offering hosting of their smart apps for those who do not wish, or do not feel competent, to install the apps on their own systems. Or indeed monetise them just because they can.

The process currently isn’t as elegant as it could be, and I assume it will be considerably improved, but once you have an app sitting on a server somewhere it only takes a matter of minutes to make it available in your location and install it via your mobile app in the already familiar fashion. Getting to the stage where you can do it that smoothly is not really that much more effort than installing GitHub integration into the IDE is. Sure the process will need to be simplified but I expect that is on the roadmap.

5 Likes

Time to jump ship to HomeKit.
They removed Robot vac and TV automations so for me nothing really left to “keep me on the hook”

Ding ding ding! Exactly!! If the new apps could be packaged as docker containers, it would become simple to run them anywhere — locally on an Rpi or PC, or in the cloud on Azure or Aws, or some 3rd party that steps up to make it click and go (and charges a nominal fee for the integration).
I really don’t want to down the road of having to deploy lambda functions; just provide me a container that I can run where I choose.

4 Likes

Can you imagine trying to troubleshoot a home automation system when each custom app is running on different cloud servers around the world? :thinking: I doubt Samsung support will offer much assistance in that scenario… I know I’d hate to be tech support if that design ever comes to fruition…

At least having the code all run on Samsung’s cloud meant that it was all or nothing.

Yes, I realize that many integrations are already cloud-2-cloud. However, most are from the big companies like Google and Amazon. Both of these recently received an overhaul by Samsung to replace the old integrations/skills that probably relied on the legacy groovy platform. The new Google and Amazon integrations appear to be an all or nothing proposition in terms of which devices these voice assistants have access to. My guess is that the new C2C integrations use the new SmartThings Platform API.

This makes me wonder about the future of the Lutron and Logitech integrations on SmartThings. Who is going to rewrite those integrations? Samsung or the other company? Who is going to host those integrations? Lots of questions…not too many answers yet.

3 Likes

Probably no one. :disappointed_relieved:

We’ve already seen the issue over the last year where the new Lutron Caseta fan switch does not work with the smartthings integration. Lutron support says it’s up to smartthings and smartthings support says it’s up to Lutron.

2 Likes

For those that want to read about the interview.

2 Likes

@jody.albritton another thing that’s missing documentation regarding adapting custom DTH to work in the new app is how to handle child devices. I.e. how to properly setup a DTH to use MCD.

3 Likes