Tbh I don’t see the alleged benefit of auto discovery and installation of bindings.
Why not just make an additional step in the startup assistant with a simple question
“What manufacturer of devices are you using?”. And some kind of searchable drop down selection thing.
Even if I am absolutely new I know that I am e.g. using shelly devices.
Or “What kind of wireless technology do you use?”.
The startup assistant would then install the bindings.
This would bring complexity down and auto discovery of things already works for some/most of the bindings.
With two or three questions it should be possible to “discover” most of the needed bindings.
And even if not - if there already is an installed binding the new user will know where to look for more bindings because the existing ones can be used as an example.
I am aware that it’s not as elegant as real auto discovery, but it should be very easy to implement and typically binding installation is only done once so I don’t think it’s worth all the extra effort.
That’s right. What that selection step could use, though, is some way to search by manufacturer and/or product name. Same applies to the web site: the addon page allows searching, but e.g. entering ‘Roborock’ yields no results, even though Roborock vacuums are supported by the miio binding.
done once when?
Quite often when a very first time user is very first trying to use openHAB for the very first time.
And what if that first time user, someone who read or saw something about openHAB and decided to give it a try successfully installs the program, gets to the main UI and spends the next hour trying to figure out how to make their smart light bulb turn on and off
don’t believe me? just check the forum. example
you know what may make OH easier to use over time is to implement OH as a Matter controller in the core of OH. Then the scope of writing a tutorial is much smaller,
Binding discovery and non-Matter devices would be not that important as they would be an “advanced use case”.
The effort for users to be able to promote OH would then be easier for grass roots marketing.
OH could highlight it’s independence and freedom from platforms that are controlled by companies
looking to monetize the platform. “Free, always been free, always will be free, free from companies looking for money from you”
To me, that sounds like an awesome marketing message!
ps… I’m looking for an easy way to get a Java implementation of a Matter controller
Yes - but there the user is already selecting bindings.
I’m suggesting to search by manufacturer or device name, see the example of @maniac103.
The binding setup is done typically once at first start (for a new user).
My argument is that if there is a smart dialogue that installs the bindings based on manufacturer, device name or technology then it’s not worth the extra effort to implement auto discovery of bindings.
Please be aware that openHAB itself is another, well… standard (at least from an outsider perspective) so even openHAB added another standard to the competition by trying to implement a more generic way to control most (if not all) smart devices.
The only “real” solution would be to force all manufacturers to use one protocol, one wireless net, one wired net and so on. But this would prevent progress and new inventions (and its not realistic anyway…)
I’m NOT advocating for ONE standard, I’m recognizing that the market is moving in a direction and OH needs an easy, good, and compelling story to keep the community engaged and the rest of the world interested in looking at OH.
TCP/IP took over the world and SNA(IBM’s networking standard) died
Sure, but that’S a complete other story.
There are hundreds of “standards” out there, as every manufacturer built its own shit (there are some exceptions, for example knx)
And there are good reasons sometimes for another standard. Matter is not.
I think my suggestion is being misconstrued.
I’m NOT suggesting that OH be MADE INTO JUST a Matter Controller.
I’m suggesting that OH incorporate a matter controller in the Core package.
Then a new user would not have to install and configure bindings for basic use cases for the time when Matter has established itself as the defacto standard for Home Automation. The OH Matter controller would interact with the rest of the OH platform just like the Zigbee, Zwave, and other bindings do.
OH would continue to offer first class support for the “much more beyond that”
Well, if (not when) Matter will some day be the “de facto standard” (I highly doubt that) there will be a matter binding, and that’s it.
There will be a matter binding sooner or later anyway, and sure enough it will have autodiscovery, but you will have to install it like other bindings too.
IF Matter becomes defacto standard, write a matter controller but don’t try to force openHAB developers to do your will. It’s nonsense to set Matter as a core preference as openHAB itself has a bus, that’s the core, Matter won’t.
and when? 2 years, 3 years from now? Will it really become the de facto standard? Even if it will become the de facto standard, the market share in the beginning will be extremely low compared to all legacy products.
Just to respond to this specifically about wizards, and sorry for the delay:
Features like “Add Equipment to Model” and “Create Equipment from Thing” are wizards of sorts, because they deal with multiple entities at once (fill out a form, have items and links created for you).
There’s a first-time user wizard, but it’s deliberately kept to the bare bones because it might annoy if you stretch it too much.
What I very much had in mind, though, is a wizard that could bootstrap your initial skeleton of a semantic model (at least the hierarchy of locations and such).
This could be offered when you go to Settings > Model and there’s nothing to show - you would have a button to start the wizard.
As for wizards that could go from discovering a Thing to using it (items, rules and all), I had this in mind too - we could import “packages” from the Marketplace which could define things, items, rule templates, main UI widgets, block libraries, etc., and then have wizards so a single form would create instances of all of these at once.
But it’s kind of off-topic from the title of this thread IMO, the discussion has shifted from “OH needs more marketing” (like, bi-weekly blog posts?) to “OH is too complex to use”.
excellent, then this part should go in a “Configuration Guide”, not in a “Gestting Started Tutorial”
or another option would be to choose the simplest rule language as an example in the Getting Started Guide and then leave other languages ffor the complete Configuration Guide.
IMHO this is questionable: step-by-step guides are great but for very beginners and for limited parts of the documentation, never seen a software (perhapes just those with billions of users like WhatsApp or Gmail) with step-by-step guide that cover EVERYTHING.
agree, what is there is there, it can be just moved/renamed to a more comprehensive “Step by Step Guides” section.
Still most easy and immediate rewards parts could be moved to a short Getting Strarted Guide.
Well, if one sees an easy Getting Started Guide using a Hue device “from zero to hero”, most users should be able do do the same but with the binding/device they have at home…
if not well, probably theiy are not a OH target user, and they should go better to Alexa/Google Home ecosystems…