In your opinion, that is. 
But I strongly disagree (IMHO, of course 
).
Please explain how having both addons and integrations in 1 list will help vs what we currently have. Please explain how having these 2 items in the same place help people learn what they need to install.
Keep in mind OP wants them as separate items in a single list, with an icon or color that differentiates the two.
1 Like
These are two completely different and disjointed points.
-
yes itās transitioning. Yes itās going to take time (hearts and minds) and no it DOES NOT happen overnightā¦ And until the itās still not for your mother.
-
when it does. Iām with P. Thereās multiple technical reasons to NOT do this that are actually protecting the end user from stupid. Mostly because powerful systems are complex. Some of this stuff (frigate) will NEVER be in the realm of end use land - nor should it be.
ā¦Else you risk becoming SmartThings. If you know anyone who was in that platform (I started there in 2012) they will tell you horror story after horror story about āmaking things easierā that have IMHO absolutely crushed the original value prop of SmartThings and now I wouldnāt recommend it to my mom or anyone else for that matter.
Simpler is not always better. In that case it pays to listen to those who went before. And yes HA tried to simplify this before and it got worse. Donāt want to Smartthings the platform thanks.
1 Like
@Boilerplate4u, @flotteemse and @santaklon ā
Youāve mentioned that HA has a long-term goal of being more mainstream and accessible and that it is failing to do so because of the techy people getting in the way. As a means of proving your point, you mentioned the complexity of showing the Addons to your mother.
This example you provided demonstrates a fundamental misunderstanding of what it means for something to become available to the masses. It does NOT mean that somehow, a non-techy user, can suddenly do techy stuff. It DOES mean that they can do the basics.
Take a Roku for example. That is something we can all agree is available and usable by the masses. My grandmother can turn on the TV, turn up the volume, select an application, and choose a movie. However, the is no way on earth she could have set up the TV, connected the Roku, downloaded the necessary apps, organized the home screen, and logged into the necessary services. Heck this is even true of some of my peers.
Setup and configuration of complex things will always require some level of knowledge of the system. Donāt expect your mom to do addons and integrations. Help her set it up and then let her use it.
If you think this understanding is errant, please provide an example of a platform that does allow such a massive amount of extendibility and virtually limitless possibilities that your mother is able to readily use. If that actually exists, perhaps we can learn from its design.
2 Likes
No!
That is not what I said. I mentioned: ā By having them in the same place (e.g. one search result but a pop-up with the two different types with links to descriptions) would make it less confusing.āā¦ā¦
None of that is in the WTH.
Well, as always, it all comes down to how things are implemented, right?
You get out what you put in so, yeah, donāt ever let inexperienced programmers take charge of UI/UX development! Ever! 
Iāve worked with complex industrial manufacturing plants where the process control is far more complicated than HA but still easier to understand. The complex settings are smart enough to handle themselves, making it simple for the operator to work with.
Itās not rocket science; itās more about adopting a way of working where developers focus on simplifying complex tasks for the end user. Sure, it might not be the most exciting thing to do, but itās absolutely necessary. And honestly, any real change here has to come from the top.
In the end, itās more about changing mindsets than solving technical challengesā¦
Iāve been working in that industry for close to 20 years now. What software are you talking about specifically, Iād love to know. I havenāt found a paid software that makes any of this easy yet. Every single one requires training at a certain level. PLCs, CMMs, CNCs, Robotics. FYI this is what I currently do for a living, integrating systems into assembly lines so that a few high school education level operators can run the entire production line.
1 Like
No. Sorry for not writing the suggested solution right away. My WTH is quite simple, that I do not see the meaningā¦ I do not see it as a requirement anywhere, that you have a solution for your WTH. I just reacted, as you wrote something which I did not say.
And yearh. This is my last message in this thread.
Your example of Roku is good and you are right, not everybody has to be able to do absolutely everything for a product to be succesfull. This works, because a large majority of non-techie users born after ~1980 who have regular interactions with computers and/or smartphones probably can set up a roku within a couple of minutes for their grandma to use.
This can however absolutely not be said for HA, and with the mindset iām often getting in this forum, thats not ever gonna change.
The question really is: Should HA become a tool for the masses (then it needs to adapt a lot, being better & more appealing than Hue, Aqara etc.) or should it become an industry tool that you pay someone to install at your home (then it needs to adapt a LOT, because then you are going up against e.g. Lutron, Crestron, Siemens, ABB).
As it stands now I see HA somewhere inbetween, where it is a fun hobby for us nerds but neither appeals to the masses, nor is remotely secure/reliable/scalable enought for professional installers to consider it an option.
We donāt have to discuss this here however, since that decision is made (or not made) by Nabu Casa and the HA project is either gonna sink or swim, no matter our personal opinions.
Anyway, thank you @flotteemse for bringin up the point & thank you all for weighing in. For me personally this topic is at its end.
I think this dicussion has - once again - shown that the user base of HA is extremely divers, ranging from people who just want to plug stuff in and have it work, to people who are deep into tech and love to tinker around with it. I think the biggest challenge for Nabu Casa & the HA project will be to figure out where in this fast moving market they want to position HA and whether this choice will be financially sustainable in the long run.
Iām here for the ride, but I do have a clear opinion on it - even if it is an unpopular one.
1 Like
100% right petro
Not only failed with Zwave
Also with zigbee. People have installed both ZHA and then Decond or Zigbee2mqtt and then nothing works or it works randomly because the first Addon to grab the USB stick wins and the others fail.
There is no way around it. You have to learn what Zigbee2mqtt and Deconz are and that they are stand alone applications.
If some wizard installed the MQTT addon when I add an integration that needs MQTT my system would break down because I run MQTT on another machine than Home Assistant.
It is so essential that users learn from the beginning the difference between integration and addon. And also learn that addon is an option and that it makes perfect sense to run the 3rd party application on a different machine.
2 Likes
In general I understand the intension with this WTH.
The problem is really that the reality is not that you install an integration and an addon that have the same name. It is not reality. And because it is not reality - the proposal is going to cause more confusion among beginners than it will help them. I honestly think most developers here deep in their heart want to make HA easier for beginners. And that also matches reality. Look at the last two years.
We have added UI for Automations, scripts, helpers (while still maintaining YAML for the advanced use). We have much more beginner friendly dashboards. It is going in the right direction.
Look at my reality.
I run zigbee2mqtt on a different computer. I run deconz on a different computer. I run Mosquitto MQTT broker on a different computer. I run Frigate on a different computer.
If a beginner wants to start with HA on a Raspberry Pi 4. And he wants to run 4 high resolution webcams in Frigate. Good luck getting Frigate to run in parallel with HA and a lot of other stuff on the Raspberry Pi 4. Running such a CPU intensive task on a small machine will not work. And not UI simplifications will help that.
If a beginner wants to run Zigbee and starts with ZHA. And then he finds out that his devices are not supported by ZHA but it works with Zigbee2mqtt so he installs that. And for that he would find two zigbee2mqtt things listed. Which one to install. Oh?!! Both. Ah. But no it does not work because ZHA steals his USB stick access so he needs to disable this first. And what if he also tried deconz?
Users have to quickly understand that Addons are an option to run 3rd party software and that running multiple of these can conflict or drain HA of resources. Or with e.g. Frigate you would want to run HA/Frigate on a beefy computer with much more horsepower. You cannot solve the problem through UI.
I honestly think listing Addons and Integrations on the same list would be much more confusing and right out damaging to beginners. It is because I care for beginners I find this idea bad. It will not work well.
2 Likes
Okay, but Iām talking about something totally different like standardized software solutions for central process control in large industrial plants, or as itās called nowdays Manufacturing Execution Systems (MES). Itās not just about trying to put together individual parts like the ones you mentioned below, which are more like how HA works today.
When I started working in the pharmaceutical manufacturing industry about 35-30 years ago, we often used Fujitsu PCS/SCADA. The operator interfaces, which used text terminals, were pretty awful back then. So, the company I worked for and I came up with the idea to develop a user-friendly, graphical operator interface using Win3.11(!), which became really popular in the industry. Later, Fujitsu bought it and offered it as an add-on service.
Todayās MES solutions from companies like Kƶrber, Siemens, and Critical Manufacturing are a whole different ballgame compared to those early days. These are mostly development systems with a foundation kind of like HA, using ābuilding blocksā (integrations/add-ons). But they really stand out when it comes to the human-machine interface. Itās not that theyāre more complicated, but theyāre designed around the actuall user process, not just a specific technical function like HA.
Just as a side note, the pharmaceutical industry has strict FDA/EMA regulations for quality control and traceability. Because of this, all systems need to be very reliable and user-friendly. If something goes wrong, you must be able to track every step of the production process. To further minimize errors, pre-production facilities are always set up before full production begins.
Here is the list of acronyms you mentioned in the previous post:
- PLC = A programmable controller with I/O devices for autonomous automation (small to midsize)
- CNC = A programmable machine for cutting, drilling etc.
- CMM = A measuring device
Yes I work for a medical company, however the distinction here is that the operators who runs these machines are still not using them the way youāre describing. They are not the people who are adding integrations/addons and programming the machines. Operators are just the people who use the end product in a fully automated state. In regards to HA, this would be equivalent to someone just using the overview page without editing it.
It would be advantageous for Home Automation suites to see how the manufacturing industry handles these kinds of issues. Many companies have came up with solutions that correlate to issues that HA currently has, the hurdle would be getting your hands on the software to play around.
Lastly, a large reason these manufactures have incentive to push for these types of designs is specifically to conform with the medical industry. Manufacturers push to be Part 11 compliant in order to sell to these medical companies. This forces them to add things like RBAC, audit trails, etc. None of that is present in the Home Automation industry.
Okay, but how the operator works is really beside the point (maybe read my text carefully again, and youāll see why). Also, Iām not sure why youāre bringing up things like CMM and CNC when youāre working in the pharmaceutical industry.
Because we use cncās to create parts, and cmms to measure them. All built into the assembly lines.
To correlate this with HA, HA would be the thing that manages the lines (PLC, or some custom software that connects to PLCs) where CNCs, CMMs, and robots are the devices connected to it.
I think the valid point here might be, that Add-On and Integrations are too close from their wording based meaning.
So if Add-Ons would already distinct them more clearly (by the name, by the description in the UI WITHOUT the need to search in the docs) from what an integration is: We might have a lot less confusion with beginners.
There is a reason that HACS has this funny āThere are no AddOns in HACSā statement in the UI ā¦
But I also think that they have to be separate lists. Cause otherwise it would be even more difficult to keep them apart.
I mean if we already mix lists: Why keeping integrations, Devices and Entities separate?
They are also at least as close connected as Add-Ons to integrations.
So yeah, let all be one big list.
(Just to avoid angry responses, that was a jokeā¦)
1 Like
This is a legitimate road to a solution. New names. Perhaps the following:
- Addons ā Applications
- Integrations ā Connectors
This is not true. Devices, entities, and integrations are directly related and inseperable concepts. Integrations are what provide entities and entities make up devices. Devices cannot exist without entities and entities cannot exist without integrations.
1 Like
I think integrations is probably the best term use for services that add devices and entities to HA as they current are called. But yes addons should be called āAppsā that you install from the HA āApp Storeā. This language is common parlance for Windows Android and Apple users these days
1 Like