Nubbie. Techie. Extremely frustrated with HA :(

I tend to agree with OP on most of the points he’s making, and I come from a similar background. I think HA should be easier to approach for software professionals (or hobbyists), whom by the way are potential future HA contributors.

The YAML syntax is terrible to learn and use when you’re able to effectively write Python code. Often it’s a massive amount of config lines to maintain, when you could achieve the same with a few lines of Python. It’s frustrating. I won’t go in to the details that are wrong. I just think YAML isn’t the way to go for coders. It’s not possible to fix it.

On the other hand, like people have pointed out in the thread, YAML just isn’t easy enough for the average Joe either, “whom just want the box”.

So the way I would improve HA UX would be:

  • Coders: Officially encourage the use of AppDaemon. It works great and saves you a lot of frustration (I would rather have skipped learning YAML altogether)
  • Average Joe: Replace YAML with something easier that’s UI based, even if it means less features overall.
1 Like

It’s really just the automations in YAML that are a pain I think. Offering less features would put us in the arena with the “box” solutions. I think HA needs to continue in the direction it’s going. I would though like to see an integrated YAML editor of some type. And an automations editor (which I think is being worked on). YAML may not be the best solution, but it’s better than XML or trying to setup python nested objects (at least for those of us without a warped sense of fun).

3 Likes

I’m a noob. I have no hardware nor have I installed the product yet. I have been trying to figure out the architecture the last few days and not gotten to the config. I think having to write any code is a deal breaker for me. There is way too much to learn about infrastructure and what works with what. Throw a new language on top, it’s to much time for me right now.

Point of importance. This was never about free. It was about security, being able to run local. Not sure that any other solution offers this.

I think a GUI is required to broaden the products market. I realize that it will take some time to reach that stage and hope that this can be a goal of 1.0 or no later than 2.0.

I am a fan of Pareto’s Principle (80\20 rule). Some one noted that there are probably a small amount of devices (20%) which many many people (80%) use, and that for these it should be possible to build some standard scripts (functional, maybe not optimal). I have been involved with OpenWrt\LEDE and they use the Table of Hardware tool to help users find compatible routers. I think this concept would be very helpful to the HA community.

The table consists of columns of features, one of which is a link to a related device page (and another to manage the tables entries). This page includes links to the device, photos and relevant config info. In HA’s case this could contain YAML config files that one could copy and paste as a starting point. It also could contain info about the devices compatibility with the various protocols, hubs, installation notes, aesthetics, links back to the mfg docu, etc. I suggest this be revision specific for each device.

Regarding the Average Joe comment directly above, I agree. OpenWrt has a config gui, but it only supports the basics. I am not smart enough, but would think that groups of objects can be made functional (individually) with the same or similar sets of parameters( light bulb => DeviceID, on, off, state - omit dimming for example), and that to start you could pick a handful of object groups to focus on (lights, switches, sockets) which are present in most every setup.

[quote=“turboc, post:42, topic:20946”]
Offering less features would put us in the arena with the “box” solutions.
[/quote]Yes, but I see this as a good thing. This opens up an avenue to users (like myself) to get in. I would not have used the trunk version of OpenWrt as it had no GUI, but the release versions do. That got me in the door. I have since learned to do some things in config files

.

For basic things like turning on lights, managing switches, and items like that. HA is basically where you want it. The components page lists all the components HA works with. Each of these has sample YAML scripts ( usually 4 or 5 lines ) that are appended to the configuration.yaml file. Then everything else is handled in the gui. Where the challenge with YAML comes in, is when you want to write complex automations. If I want to turn on the outdoor lights after sunset, on weekdays, but only if it’s cloudy outside. When multiple conditions start to be included, YAML gets complicated. Just to turn on the lights at sunset, is a simple 3 or 4 line YAML automation, and there are many examples of that on the forum.
I believe HA already provides it’s target community with 80% of what they want. It’s probably never going to get better than that, because the target community always wants more. The question comes down to defining the target community. Based on your description of yourself, I would say you could be part of the target community. It seems you are not afraid to roll up your sleeves and dig into things. That is the community of people that HA is best suited for. If you want an out of the box solution, that does 80% of what you want today, but will still only do that same 80% six months from now, then a box solution may be for you. I like the flexibility that HA provides me. It is growing with my house. I get new ideas of things to automation by being a part of this community, and they push me to learn more.

OK, I actually had seen the config pages early on and forgot about them. I consider them a bit thin. Scrolling through things like senors it’s hard to figure what most do. Needs some type of Application information. As noted above, could benefit from more info IMHO.

I did some more looking at the site as opposed to forum. About the only thing I know from the home page is that I can store my data locally, which is important to me, indeed why I am here.

I think that the Getting Started section is a problem, indeed a problem with many web sites today. It assumes I am sold.

I want to learn more about the architecture and the way things fit together before I start loading software. How about some discussion regarding the different groups of tech, like Z-Wave is better for this or that and Zigbee is better for that or this, and how they work with HA (what does what). Clarify terms. How about pros and cons of Pi vs NAS vs Virtual vs PC? What about cameras and video, what resources come form where? Does this impact hardware selection? There is no suggested hardware, just a list of platforms. I see no “Security” info on any of the home page links or documentation. No discussion here about installing mobile devices. I have a dozen hours searching and reading forum messages for stuff I think should be clear and upfront.

Please, I am not trying to be negative, but as the title says, …Extremely frustrated…

Is there a water sensor (leak detection)?

1 Like

You make some good points. As a user contributed open source project, documentation is one of the things that suffers. It seems that no one likes to do documentation. :slight_smile: I know I don’t. This is where I normally say you have a job and in truth, we could benefit from any help we could get on the documentation side of things. Documentation seems to be one of those things, that the less you know about the topic, the better you can do it because it has to be written in such a way that it explains everything to you.
You may have an answer to your question at the end, but Yes and No. The thing is, it’s a sensor. HA doesn’t have a water sensor, other vendors make a water sensor and it can be integrated into HA. That’s part of the beauty of HA and why and architecture diagram is difficult to achieve. The architecture is kind of like this. You have a HA server and you have a bunch of other devices that other manufacturers make that talk to it via a bunch of different protocols. It’s kind of like your brain. You have a brain and I have 5 senses, that tell me all sorts of things. Those are the well defined fairly easy to integrate things. Then there is this vague 6th sense that wonders in every once in a while that I have to figure out how to deal with. Those are the items that are more or less well defined in the system. They make up about 10% of what we have, and if you are using them take up about 90% of your time to figure out.
I won’t lie to you. HA is not for the faint of heart. It takes time and thought to make it work well. And just when everything is working well, a new release comes out that changes all your zwave device numbers, and you spend a long rainy weekend fixing things. The good news is, you know it’s coming before you decide to install the release so you are in control of your system, unlike some systems like Wink that drop updates on you and you have to figure out what happened when things around your house start behaving differently.

But back to the documentation. It does need work. I am sure that someone can help point you in the right direction to help improve it, if you are willing to help and take up at least part of that challenge.

I think the documentation is descent for an open source project. I first started with home automation using a product that had very extensive and detailed documentation and I recall being very lost for awhile.

I spent a lot of time on cocoontech learning all the different things. HASS is only one small and not even mandatory part of home automation. There are a billion things out there that are compatible with HASS that aren’t even mentioned on the website.

To me it seems like the issue is more “home automation is overwhelming” in general and not a HASS problem per say.

1 Like

Documentation seems to be one of those things, that the less you know about the topic, the better you can do it because it has to be written in such a way that it explains everything to you.

This. Definitely. When you’re too closely involved, you miss the things that a newbie would have trouble with because it seems obvious to you.

3 Likes

Deleted by author

1 Like

I think my thoughts can be summarized thusly:

  1. HA sucks in so many ways, many listed in this thread.
  2. HA is by far the best system available, for someone who wants to use a diverse set of devices, and is not willing to rely on a 3rd party cloud dependent solution.

@anderson110 ok then what home automation system would you suggest as being better? (And by better I mean has the breadth of components and flexibiliy of custom components)

I also get frustrated with HA at times, specifically the breaking changes, but after having played around with most of the other platforms out there (from Smartthings to Openhab and domoticz) I am backing HA with my time and learning curve commitments.

Bottom line is if you want a point and click solution, you will be limited to the likes of ST, but if you want flexibility and diversity, you going to have to roll up your sleeves and get involved…

1 Like

There are basically 2 issues mentioned: 1: YAML sucks, 2: component do not work,
Point 2 is not the case for most of the components, which is backed by several posts.

Point 1: YAML sucks for programmers. A lot of people do not care about semantics and can get around with the YAML quite well. But, agreed, a more point and click solution would be great.
So, your many ways are not that many in fact, but maybe you can give some more thoughts of why you think HA sucks in many ways.
Also, what is the best system available in your opinion?

There are many platforms available, if you do not like HA but favor other platforms as superior, that fine :slight_smile:

Might be wrong but I think you, and @Romkabouter, took @anderson110’s post the wrong way. I took it as he was being positive about HA by saying it’s the best thing out there but it does have many flaws.

This is true, it’s a long way from being a polished commercial product. But, because it isn’t, what it can do is only limited by your imagination/skill level and it continues to grow on a daily basis.

4 Likes

Try reading point 2 again.

I think indeed I misread point 2 :smiley:
Sorry about that :wink:

I know that the dutch Homey uses an old version of HA (there were some rants about it somewhere), they look like having a nice UI to add devices and flows (automations)

Having such function in HA would make it much more useable for non techies

When HA worked I thought it was great and its breadth outstrips all the others but when an upgrade trashed the lot I moved on and periodically return to see if things have improved. As an experiment I have two Raspi and I tried to put HA on one and OpenHAB on the other following the basic installs. Suffice to say neither worked out of the box and neither works now after wasting too much time on issues that shouldn’t be. For example if Etcher won’t work with the HA image why recommend it? Why put up a half-tested image as the latest release? (Testing of this image has not been as extensive as the previous images but has been running well on a Raspberry Pi 3B+ for a few hour, as of publishing this image.) It doesn’t work for me!
Back to NodeRed!

I’d second the OP to be honest - I understand the driver behind making a self contained ‘appliance’ style image, otoh you could make a reasonable argument that improving the configuration of the software would make actually help a greater number of people [the type of folk for whom hass.io makes things easier are going to be hit with a huge wall once they actually try to configure it to do anything slightly non-trivial].

The debug code on a lot of the components seems to be poor to non-existent, and so I’m having to put in debug code myself in order to diagnose issues with my configuration.

One of my reasons for going with home-assistant was the language choice, and the idea that if I needed addons I could always write them myself in Python. I just didn’t realise that it would be necessary to get core functionality out of the thing.

I’m generally moving away from trying to write actions in home-assistant, and using it as data collector, with my flows in custom code or node-red.

If you want to do this in Python, check out appdaemon.

I use Home Assistant purely to hold my ‘states’ and run all logic from Node Red. Much easier to make changes without reboots/restarts and I have debugging tools right there in front of me.

I’m doing something similar - using HomeAssistant merely to discover states, and hold current states [longer term states I’m logging in influxdb] and then running logic from node red.

If there was something like a standalong appdaemon-a-like that could trigger off mqtt, I would probably go straight from HomeAssistant to mqtt and do everything in that plus node-red.