Initial choices for a new user

  • Platform information:
    • Hardware: Intel i5, 8 GB ram, 250 GB SSD. (Testing on a Microsoft Surface Pro 4)
    • OS: Windows 10 Pro 64-bit (Version 2004)
    • Java Runtime Environment: ZULU Java 8
    • openHAB: Version 2.5.6
    • Jython: Version 2.7.0

File Structure:

I think I reported my file structure incorrectly in a previous post. I have really struggled with getting the file structure correct since I see differences when looking around. It would be nice if the references to the file structure would show the full file path and be clear if Windows 10 or another operating system. I may have this install incorrect but will try to adjust as needed when I am sure what the intended file structure is.

I am a new (or at least hope to be) user of openHAB2. I have no rules to convert so that is not an issue here. I have used VB.NET for my more recent home automation efforts. Several years ago I used HomeSeer and prior to that I used a system called Homebase. I wish VB.NET was an option for scripting in openHAB2 but as near as I can tell from what I have read apparently it is not.

I would like to avoid spending time learning things that will be going away since I assume I would want to move to openHAB3 as soon as that is an option.

I can see there are a lot of very experienced and knowledgeable people here and I would be appreciative of any thoughts or words of wisdom any of you may have on how best to proceed in a way that will make things easier down the road. I feel somewhat like a first grader in a room full of engineers.

Here are some initial thoughts, some of which may be way off base.

Add-ons - Paper UI
Bindings - Paper UI
Things - Paper UI
Channels - Paper UI
Items - Files
Groups (if used) - Files
Rules - JSR223 Scripting - Jython/Python

Should I assume it is too early to try openHAB3 ?

If so, any ideas on how long before openHAB3 might be available?


It appears you background is just with Windows systems. OpenHAB is designed primarily for UNIX-like systems but Java permits Windows too.

Channels are defined by the Things. For my Z-Wave devices I prefer to use HABmin to discover and add Things. In my opinion it is best to add Things within OpenHAB. Using text files for Things can cause a lot of issues.

My understanding is that OH3 is not yet usable.

@5iver or @rlkoshak can comment on the rules. I have not yet moved to that newer system but it is the powerful way of the future.

I know that there was a nice table overview in this forum comparing the recommended ways of configuration for different openhab version, maybe even including v3.
A fellow pointed it out to me to show that text based configuration is not always the best option - but please, don’t ask me where to find it again :sweat_smile:

Edit: found it:
(Chapter Versatility)

The directory names are case sensitive and must be in lower case.

Ignore all documentation for scripted automation except for this…

For the directory structure, see… Other than the root directory, the directory structure is the same for everyone.

VB.NET is absolutely not currently compatible for use in OH with the new rule engine and scripted automation. Currently, only scripting languages that implement a scripting engine using the javax.script API are compatible.

OH 3.0 will be a hot mess, particularly for the rules DSL. Starting out in OH using scripted automation, Jython, and the helper libraries has MANY benefits, and it currently looks like there will only be minor changes needed to use these scripts in OH 3.0, such as changing the directory structure.

Until available in your UI, you can easily install Jython and the helper libraries using this…

Just be aware that if using Windows, the scripts will not run after OH is started. This will be corrected in OH 3.0. I have another jar that will correct it in OH 2.5.x…


Items are best in text files until the REST API is built out and then hooked up in the new UI (OH 3.0).

Let OH create your Things. You may get some misleading feedback about this regarding backups, but ignore it, since the Things are all in the jsondb (text files) and very easy to backup and even edit. With a manual installation, like is done for Windows, backing up OH is as simple as copying a directory. I still prefer a manual installation of OH over a repository installation, and I use Fedora.

I suggest not using OH on Windows and having it bogged down with all the bloat. If OH is not run on dedicated hardware, it is very easy to slow it down. When it comes to things like turning the lights on, every cycle counts in order to keep you from standing in the dark… even if it is for just a fraction of a second :slightly_smiling_face:.

Take a look running openHABian on a Pi. This will help a bit with any Linux learning curve.

OH 3.0 is available for testing, but it is much too early for production use, especially for a beginner.

I’m not sure why you are fiddling with directory structures.
I just use this page

to select a stable download for Windows, then unzip it. Directories done.

If you want to host on a Windows box, that’s fine. Some of us have something suitable already in use for some other purpose, sometimes an old laptop is available for those first investigative steps. openHAB doesn’t mind.

DO NOT select Simple Mode at first bootup. It’s tempting, but becomes a complete pain as soon as you want configure to something for yourself.

I would recommend start with Demo, have a good play with its innards, then trash that whole install and start afresh.

Note that Paper UI * HABmin are gone in OH3 replaced by a new UI. The chart may be a good guide for OH2 but I suspect they are planning even better things for OH3.

EDIT I just found the OH3 page and they currently have the same, obviously incorrect, table.

1 Like