Then you should try OpenHAB.
It will teach you how wonderful Home Assist is. Trust me: I used openHAB for 3 years and achieved more with Home Assistant in 3 weeks.
3 Likes
They appear first (on top), in bold, with a “Discovered” title.
If they were placed on the bottom and you were using a phone, you would first have to scroll the list to the bottom to see them.
Placing them on a separate page would mean the moment you finished configuring an auto-discovered integration, it would disappear from the “Discovered” page leaving the new user (the kind who is “unnerved” by the current system) unnerved by its sudden disappearance (it was automatically relocated to the proposed “Installed” page but they probably wouldn’t know that until they looked there).
The current system (auto-discovered integrations appear first and in bold with a self-explanatory title) has served countless users well but could be improved by HeyImAlex’s suggestion to initially show the button as “Install” as opposed to “Configure”.
It does notify one in Notifications but I feel they should be on another pull-down page…
I do realise that Home Assistant is very powerful but it’s a pain in the bottom at times
Put a note that the discovered device/entity has now been installed and is now in the Integrations page along with the rest…
Just a quick question off subject: Can one log into the ‘Owners’ account from two devices, PC and Phone, at the same time?
If you have to post popup messages explaining what just happened then it’s often an indication that the UX needs improvement.
People deal with it and discovery is usually a one time thing. That doesn’t mean it’s good UX. It also caught me by surprise the first time I saw it on my dev instance (HA doesn’t discover things on my production system as I disabled all discovery components). ‘Discovered’ is not self explanatory at all, as it doesn’t give any reference to installation status. Something like ‘Discovered and ready to install’ would be a lot better, along with the install button.
Doesn’t have to be a separate page, can be a separate section on the integration page. The integration thumbnail would change sections once it is installed.
I can’t speak for other English language speakers but for me it means Home Assistant discovered the displayed device.
Changing the newly discovered device’s “Configure” button to “Install” would clarify the intended action. I agree it would be an improvement.
The existing presentation has “implied” sections (where the “implied” aspect might not be equally obvious to everyone so it’s not beyond improvement).
- “Discovered” is displayed first (i.e. top) and in bold and labeled eponymously.
- “Installed” appears afterwards without bold or label.
- It works with any screen size.
Explicitly labeled sections would have to meet the same ‘any screen size’ requirement so if anyone has a proposal to achieve that, post it as a Feature Request in the Frontend GitHub repository.
It usually isn’t. Based on the screenshot you posted, it appears you never successfully installed the discovered integration so that’s why it doesn’t offer a means of deleting it (your options are to configure it properly or “Ignore” it, as offered in the menu).
True; it can be. In this particular case involving the Honeywell Lyric integration, other home automation software doesn’t fare any better.
Here’s how to install it in Hubitat. To be fair, it’s a custom integration (custom “handler”) so there’s an initial installation step (just like there is for custom integrations in Home Assistant) but afterwards it’s the same as in Home Assistant and requires creating/entering a Honeywell developer account, etc.
As far as I can tell, OpenHAB doesn’t have a binding (OpenHAB’s term for Home Assistant’s ‘integration’) for Honeywell Lyric (feel free to correct me). However it does list discovered devices separately (in an “Inbox”) from installed ones. Here’s the documentation for the process of installing Things (OpenHAB’s term for Home Assistant’s ‘device’).
Yes, but what does discover mean ? Okay, HA found a new device - and now what ? This extra step of having to install a discovered device is a design limitation in HA (I assume due to the way their config flow works, I didn’t check the source). There’s no inherent technical reason for this additional step. And in modern mass consumer facing products, this extra step is largely gone. A discovered device is directly usable. Apple is specifically good at that, Windows is more or less there too now. So it is understandable for a new user without prior exposure to HA to assume that a discovered device is immediately usable. Especially if no further mention is made that it actually needs explicit installation.
Sure, but UI layouting issues are always solvable and are part of making a better UX. I don’t have any specific proposals though, I suck at UI design. As a software engineer, my UIs are unusable programmer art that everybody hates 
This has served me well:
- If I want it, I click Configure. If I get stuck I read its documentation.
- If I don’t want it, I click Ignore.
FWIW, I would not want it to automatically install itself (in situations where no user configuration is explicitly required) for the simple reason that some devices can be represented by more than one integration. For example, if the device is also Homekit compatible, the Homekit Controller integration will be presented in addition to the device-specific integration. I don’t want new entities registered without my awareness/consent.
Oh I agree, I wouldn’t want an auto install either. Especially since the HA technical architecture doesn’t allow for a flexible undo or integration sharing for devices. But I wouldn’t be surprised if new users assume this to be the case, considering it’s pretty widespread in tech outside of HA and the current wording around discovery and configuration is pretty confusing.
When I was first confronted with this discovery thing, my first thought was, ‘ah crap, why did it just install an integration I didn’t ask for’… before realizing it didn’t actually install it.
Even so, the way we see all this, as long time HA users, developers, technically minded people, we all suffer from familiarity bias. And since HA’s roadmap seems dead set on the mass consumer market (I think you know what I personally think of this, but eh), the UX has to follow if we want to avoid frustration and unhappy new users. Point in case, the OP.
FWIW, waaay back when the current presentation of devices was first implemented, I had suggested it be presented as a list, like many other things in Home Assistant. One can sort on any column including, for example, the integration’s status (discovered/installed/etc). The idea gained zero traction (and my discussion with a developer failed, completely, to gain a convert to the suggestion).
Since then (and it was a long time ago) I haven’t seen any Feature Requests or WTHs asking for improvements to its presentation. Seems like whatever friction new users may encounter with it hasn’t been sufficient to raise it as an issue. Until now that is but more than one voice (or even a dozen) is often needed before the idea attracts someone to implement it.
I don’t know, I would have preferred a list too for this particular case, but I think that’s developer bias. People generally don’t like long lists (look at how much people hate the long automation list, and rightfully so). This tile layout thing seems quite popular these days to make things look attractive (you can add brand logos and stuff), so I don’t think it’s a bad choice per se. UI design that conveys complex concepts in a easy way to new and less technical users is hard, especially if you want to keep it functional for power users. But sometimes just changing terminology a little helps to bring the message across.
I’d prefer they get rid of the UI altogether and let us configure everything programmatically in code with NodeJS and C++ APIs, but well, I guess I’m a minority here 
1 Like
I managed to Install/Configure my three Honeywell TR6 Thremestates.
I already had a developers account with Honeywell/Resideo - I believe Honeywell created Resideo so they can, in the future, remove support for these Thermostats without damaging the Honeywell name too much. I had all the developers’ code string and the secret string to enter into the HA configuration box.
This kept bouncing to a web page informing me I was in the wrong place; see what I put in the first post.
I look up the Honeywell github info and was told I need to create a new developers app but this time without any special characters in the name - which I did - new code and secret strings.
I then used the original return web address and all worked.
Then I had to remove all the HomeHub integrations that I was using to get around the Honeywell HA integration problems.
Still a lot of learning to do but it is still not easy…
As for the HA finding entities, I still think they should be on another page with a brief explanation of what needs to be done to install what has been found.
As for the English: “Configure” implies you have it and it needs to now be “configured”. Install implies I’ve found something that might need some software to be installed to control it and after the installation, one might need to configure it to fit your needs…
Thanks for all the comments and help, I’m now going into an even more demanding issue and that is configuring the entities behind a Broadlink Universal IR/RF controller (BTW, it’s a great device).
IR is turning on my front room small alarm and the IR is controlling my garage doors and Kitchen, three, blinds. With Amazon Alexa it’s so easy - I don’t imagine it going to be so easy in HA, I hope I’m wrong…
I may be confused…
But what you wrote in that paragraph is the way that it is now so I think it’s you that isn’t understanding.
All of the core built-in integrations are already installed so no need to “install” anything.
once HA discovers a device/service on your network it tells you you need to configure it to use it. Using the already installed software.
So, it’s just like you said - “Configure” implies you have it and it needs to now be “configured”.
The software to control it is already installed. Or it wouldn’t have discovered it in the first place.
To answer what I believe is the real question here… “Why is everything so hard to do” with regard to having to create developer accounts, configure callback URLs, etc. It’s because so many vendors are actively against other systems integrating with their tools unless they get paid $$$. They want everything to go through their cloud services, and want that locked behind developer accounts and API keys to they can limit how many calls you can make and sell application plans where you have to pay to do more than x api calls. Another one of the downsides to many cloud products.
It’s theoretically possible that Nabu Casa could implement standard Home Assistant “apps” for some of these cloud services, but based on the install count it could quickly become cost-prohibitive and I don’t think NC (or the community) wants to get into the business of charging for HA Cloud plans that are priced depending on how many integrations you have in order to fund all these cloud connectors. So unfortunately, the burden falls on the end user to create an “app” in a vendors developer portal for the HA connection.
Yet another reason I try to avoid cloud-based devices wherever possible.
2 Likes
For instance: My Sony TV was found by HA. I opened the Integration page and there it was saying:
Discovered
Living Room
Television
HomeKit Control
I press on Configure and it brings up a configuration box that says it’s configuring - it carries on for 10 minutes until one realises it’s not going to configure. No cancel button, one has to reload the page to get the box to disappear.
One goes back to the integration page and goes to ‘Documents’ where one is told the following:
"
The HomeKit controller integration allows you to connect accessories with the “Works with HomeKit” logo to Home Assistant. This integration should not be confused with the HomeKit integration, which allows you to control Home Assistant devices via HomeKit.
"
One then installs the HomeKit controller and NOT the HomeKit integration as instructed.
One then runs the HomeKit Control and finds that it finds the TV and the TV pops up two four-digit numbers. One puts them into the integration only to be told they are wrong. One does it again without the space - No - wrong again.
Head scratching - then on notices the two four digit numbers have to go in with this format:
XXX-XX-XXX
Then it all works except the HomeKit Control is now installed showing the Sony TV but the first found Sony TV configuration box is still there. One tries to configure it again but then gets another error as it’s already configured.
Okay: just press ignore, nope, that is just ignored, and nothing happens.
Now after looking on the interweb for 1/2 hr to see how to remove the found TV configuration box one gives up and moves on to another page - one’s return to the Integration page one finds it’s gone on its own! Didn’t even say goodbye…
This is what I mean by everything is a pain in the Bum to get going, once it’s going I believe it’s rock solid but until then 
Stay Cool,
Bob…
The next project is to get the IR/RF devices hanging of the end of a BroadLink universal controller working via HA.
Wish me luck…
Regards,
Bob…