I’v just recently been looking into Lovelace, and I have to say it looks amazing! However it made me curious of the plans for Lovelace and the current polymer ui. Is Lovelace going to replace the old UI and is the old interface therefore deprecated? Is it worth investing time in setting up the Lovelace UI next to the old UI or will the format of the configuration change in the future?
Futhermore, are there any ideas to make the interface easily configurable via the UI, as opposed to editing all these yaml files? In my view this makes home assistant not very user friendly, and it is the only thing I think home assistant needs to be used by everybody.
Personally I would go for it with Lovelace! I really love how much flexibility it has over the old UI. It isn’t really that hard compared to the old UI once you get the hang of it. Try it and see how it goes and check out the love lace migration add-on for HassIO.
I’m at the point now where I have everything setup in lovelace the way I like and I rarely revert back to the old UI.
I’d like to be able to remove all the dozens of redundant groups I had set up. Also I believe there is a performance hit at the moment when lovelace is loading due to the old UI system. So if they removed the old UI from a scheduled version of home assistant (6 months time for example) then I wouldn’t mind.
I’m in @jdbrookes position. In fact, i may go through and remove all traces of my old ui configuration. Lovelace is just so much better in comparison to the old UI. No need for an unnecessary amount of organization in grouping.
As for adjusting the UI in the UI. This will be the last piece of the puzzle but I wouldn’t expect this functionality anytime soon. HA just started to add component configuration through the UI.
I recall a podcast where Paulus mused about the need for Lovelace to eventually behave like the classic version where it automatically renders a basic UI without having to define it with yaml (i.e. derived directly from entities). Once that goal is attained, then I believe the classic Ui becomes moot and can be put on the deprecated list. Until then, the auto-generation of a UI is an important differentiator; it’s a huge time-saver, especially for new users or anyone who isn’t very particular about the UI’s appearance.
FWIW, I’ve been using another Home Automation software since 2008 (and it was released 8 years prior to that). One of its many innovations was each device (light, sensor, pump, fan, fridge, etc) had an associated UI widget. After defining all the rooms in your home and populating them with your devices, the system automatically produces a functional, web-based UI. It’s also dynamic; add a device and the UI updates instantly (no system restart required). However, that product is going on 18 years old now (its development ceased in 2006!) and the UI relies on an ActiveX plugin working exclusively with Internet Explorer. Yeah, that ship sailed away a long time ago! Anyway, the ability to auto-render a UI is a significant advantage and I look forward to the day Lovelace has it.
I think we are waiting on a few more card types before that can happen. Media_player was just added not to long ago, so we gotta be getting close to that milestone.
Thanks for you input. I’ve been looking into the UI yesterday, and at first glance it seems the lovelace interface has indeed more support for custom elements and is more flexible.
Today, as a side/learning project, I started digging into the code of the frontend, learned myself some polymer (with the help of the existing frontend) and in the end came up with this (hacky) result:
. Don’t know if it will ever make it to the main repo though.