Can we revisit the move to qt-openzwave?

Configuration Z-Wave
81

Dont have that. I have all “newer” zwave plus devices mainly included unsecure (secure makes instable…).
Newer devices do support instant status updates hence do not require polling.

I just want to stress:
I was (still partially am) a long time Vera user (200+ devices) and +12 years (!) and know about zwave pain…

QT OpenZwave Beta is not painless but at least I know where the pain is…

It would very much help me if the doctor could help me to at least frequently prescribe me some medicine (update)…

82

Frenck is joining … :slight_smile: … no he left …

83

Not even close, out of 60 devices about half of mine need polling. I don’t have any serious slow down with native z-wave, but when trying qt-openzwave everything is delayed. There just aren’t the options to make it so everything isn’t polled at the exact same time.

3 Likes
84

That discussion took place on the add-on repository, in an issue, on my personal project.

I’m sorry, but:
A) That wasn’t an issue for that add-on
B) That add-on isn’t part of the Home Assistant core project or managed by the core team. It is my personal project.
C) Was completely out of context for the repository it was in.
D) From that add-on perspective there is 0 I can do.
E) Since it is my personal project, I decide myself, if I want to create additional add-ons and put my spare time into maintaining such a thing.

I’m not sending your mailbox full of discussions and complaints about the color of the paint of the house at the end of the street either. You don’t have to agree, but it was out of place there.

I find it rather annoying, rude, and plastering the way you place it in your starting topic.

(Without reading anything else in this thread, as I kinda lost interest after reading the quoted part).

3 Likes
85

So where can we have the discussion? In that issue, I asked this question, among others, but you quickly dismissed it and closed the issue without any response.

Not trying to “flood your inbox”. We as a community are looking for a discussion about what to do. I’m sorry you lost interest in this thread from the very beginning, but I would love if you could reconsider to at least skim through it. There are a lot of good questions and concerns that are just left hanging.

7 Likes
86

Can I ask the dumb question then?

Who in the core homeassistant team can say anything (content) about the future of zwave (the flavour) and the the roadmap?

7 Likes
88

Really well put, @jwelter. This captures how I’ve been feeling lately very well. It may sound a little harsh, but we’re very frustrated and the responses we’re getting reflect very poorly on the entire group of Nabu Casa and the core open source maintainers. It’s very sad and unfortunate that we’ve ended up here.

Edit: wow, flagging the post and hiding the content illustrates the point pretty well.

14 Likes
89

Frenck,

I studied the links, the github, your response and this entire discussion. But I do not really get what you mean.

Can you please elaborate on what you mean without metafors and with regards to (qt open)zwave and homeassistant integration.

Thanks.

Edit: I really think “you” have gold in hands if this openzwave thing start to really work more well!

2 Likes
90

Yes, this is the way it goes in this project. I give up and moving on.

Thanks to all those who have helped me along the way - some amazing project ambassadors on here. Good luck and hope @balloob finds a way to manage the community as well as he manages the development process.

3 Likes
91

@paulstronaut The zwave2mqtt add-on repository is not the place to discuss “the future”. It is unrelated to OZW, OZW-QT, OZW add-on or the OZW integration as was also posted in that issue when Frenck locked it. You chose to ignore that information and instead call it rude. Now you’re here posting about wanting to start a civil discussion but don’t start off civil.

The post by @jwelter was putting insults in his “personal feedback”, so correctly hidden. I don’t understand why people want to piss off the people that are building Home Assistant. Best case nothing happens, worst case they quit? :man_shrugging: Stop it.

Open source is run by volunteers. It’s what people work on in their spare time. On top of that, Z-Wave is very complicated. We should be happy that anything is done around Z-Wave in open source.

Nabu Casa exists to run the Home Assistant project. But we don’t maintain all the drivers. We are working with volunteers as @cgarwood on making OZW work well with HA. There is not enough resources available for Nabu Casa to build out and maintain a whole Z-Wave stack.

People want to know about “the future” but when it comes to volunteers, the future is what you make of it. How you spend your time to develop things.

Home Assistant has a 1000+ integrations and it’s obviously possible to integrate a protocol in different ways, like is done with Z-Wave: You can use zwave, ozw, vera, fibaro. Think Z-Wave JS is better? Great, create a Z-Wave JS integration + add-on. The suggestion in this thread that we would block this is wrong.

Yes, it’s concerning that Fishwaldo hasn’t been active lately. Last time I checked he was doing ok but it has been a while. We can’t force him on spending his spare time on OZW.

OZW is very mature. Fishwaldo is very knowledgable about Z-Wave and also talks directly to SL (who develops Z-Wave). Our old Z-Wave integration is based on OZW via Python. Python binding is not part of OZW and is unmaintained. The new Z-Wave integration is based on OZW via MQTT. MQTT binding is part of OZW and is maintained.

We consider OZW currently our best bet for integrating directly with Z-Wave sticks. Sure there are bugs, but Z-Wave is also complicated, things will get fixed. Someone not being active for a couple of a months on a project that has been going on for 11 years (!!) can happen. Maybe there is even a pandemic going on and they have their kids home?

17 Likes
92

Bedankt @balloob
Happy to hear Fishwaldo was doing fine when you checked a while ago (hope he still is).
Agree with what you are writing. I think on all sides the communication could have been better and insults are never appropriate.
I conclude from what you are writing the plan is still to go for OZW with the help of Fishwaldo. I also understand that the timing for that has become highly uncertain but that there is no plan B.
While that is unfortunate, I have not paid someone for a service that is not being delivered, so can only be grateful for what home assistant can already do (which is pretty awesome).
I do think this type of (sometimes heated …) discussions will keep popping up as long as the uncertainty lasts.
Meanwhile I guess the official advice for newcomers will be to either user the current native z-wave integration or the OZW bèta.

93

Please do not consider zwave a Legacy protocol or Just One of the 1000ds Just because zigbee and wifi have become more populare or gartner put Them in the fading Path. There Is a lot of users and devices out there.
I am not complaining in any way on something people Is working for free and gives for free buy this discussion s a chance to highlight the situation Is not optimal and the requirement Is there.
That said Happy new year and look forward for progress.

6 Likes
95

I really appreciate you taking the time to respond, but sadly, this continues the streak of kind of a non-answer. There were some concrete questions laid out that I would love to get more info on.

  1. Why does HA recommend just a single community-maintained (yet also not, because it has an “official” add-on) for Z-Wave, but documentation for others, like ZHA, mention other integrations that may work better
  2. Where was the decision made to officially support qt-openzwave? Who was involved? How do others get involved?
  3. Was there any sort of RFC process? Maybe that’s something that HA could look into for future large changes. It has worked really well for projects like React.js
  4. Why does it seem like we’re locked in and the communciation is that this will not and can not change to a different addon/integration?
  5. What happens if the core maintainer of qt-openzwave decides they are unable to maintain the project going forward? They’re the sole owner of the repository, so we’ll have to wait for someone to step up and take it over – what happens if they can’t continue to maintain, and so on? (yes, I know other projects also have just a single maintainer – it would be nice to know what the backup plans are for all integrations, should one get a critical bug and the maintainer becomes MIA)
  6. You say that it’s up to us as volunteers to make the roadmap and build “the future”, but then go on to make it sound like only Fishwaldo has the expertise to do it. How are we supposed to feel confident jumping in to something that is so complicated that you need connections with the developer of Z-Wave (SL) to do it?

Exactly why I asked where the appropriate place would be, but was ignored and the thread was locked instead. I didn’t ignore the information I was given, but merely asked what to do instead. What I felt was rude was the line by a core maintainer, “I have no interest at this point”, then again, in this thread saying that they’ve lost interest. Where do we actually go to discuss if we can’t start a discussion? Also, I’d like to note that this was just one example. Take a read through this thread and note a forum moderator’s tone and abrasiveness as well.

Lastly, I’m sorry if anyone thinks my opening message or some contents of it were out of line. I tried to be as objective as possible with links to back up any claims. I never once pinged someone directly because I didn’t want to pick on any one person directly, nor make them feel responsible for anything. We’re all in this together and we’re looking to make it actually feel more like what we do, our issues, and our opinions matter.

14 Likes
96

You can add through the UI. Configuration → integrations → click configure on the openzwave integration, then click on add node or remove node.

This will not work for security devices. You need to go to Developer Tools / Services / ozw.add_node and add the secure: true parameter. Overall this experience is not great, since you need to subscribe to the OpenZWave topic to really see what’s going on.

The OpenZWave Windows app makes adding security devices easier. Honestly, I recommend that app to anyone that is struggling with OZW.

I use OZW because I want to move to Z-Wave+ as much as possible, including locks and newer switches like Aeotec’s Nano Switch that support security/encryption.

I’m taking a bet that OZW is the future of Z-Wave on HA.

2 Likes
97
  1. Why does HA recommend just a single community-maintained (yet also not, because it has an “official” add-on) for Z-Wave, but documentation for others, like ZHA, mention other integrations that may work better

Because nobody else has written another Z-Wave integration for Home Assistant. You can’t mention an integration that doesn’t exist. For Zigbee, people from the community wrote ZHA, others from the community wrote the Deconz integration, etc. Nobody has stepped up to write a zwave-js integration or a full python zwave library and integration, or zwave-on-go or zwave-on-language-of-the-week.

  1. Where was the decision made to officially support qt-openzwave? Who was involved? How do others get involved?

History lesson time. Once upon a time in mid 2019 someone decided to undertake rewriting python-openzwave (the wrapper the “old” component uses to interface with ozw1.4) to support openzwave 1.6. I spent multiple months trying to update the old component to work with the “new” python-openzwave but never got anywhere because with every commit the python-openwave author was making, they were changing the API, and seemed to have little interest in maintaining a stable API that we could use. I talked to balloob, we talked to Fishwaldo and some others and decided to move to qt-openzwave and an entirely new integration.

The only other Z-Wave integration that existed at the time was zwave2mqtt (which still depended on OpenZWave, so qt-openzwave was cutting out the middleman) or Sigma had some information out about a Z/IP layer but that required a special Z-Wave stick and details on it were difficult to come by. Requiring a special Z-Wave stick was a non-starter for us to put resources into, as 99% of people aren’t going to want to replace their perfectly fine Z-Wave stick.

How do others get involved? Write code, submit PRs. Participate in architecture discussions in the architecture repo. I spent some time contributing small enhancements and fixes to the “old” zwave component, and be active in the dev Discord channels.

  1. Was there any sort of RFC process? Maybe that’s something that HA could look into for future large changes. It has worked really well for projects like React.js

Check out the home-assistant/architecture repo.

  1. Why does it seem like we’re locked in and the communciation is that this will not and can not change to a different addon/integration?

Because, again, nobody has stepped up to write another integration based on a different backend. If someone writes a local Z-Wave integration based on another backend that works well then you’ll see the communication change to be similar to Zigbee with its multiple integrations.

  1. What happens if the core maintainer of qt-openzwave decides they are unable to maintain the project going forward? They’re the sole owner of the repository, so we’ll have to wait for someone to step up and take it over – what happens if they can’t continue to maintain, and so on? (yes, I know other projects also have just a single maintainer – it would be nice to know what the backup plans are for all integrations, should one get a critical bug and the maintainer becomes MIA)

Historically, someone that contributed to the HA integration would fork the underlying library and HA would start using that fork until the owner of the original repo came back or someone else took over.

  1. You say that it’s up to us as volunteers to make the roadmap and build “the future”, but then go on to make it sound like only Fishwaldo has the expertise to do it. How are we supposed to feel confident jumping in to something that is so complicated that you need connections with the developer of Z-Wave (SL) to do it?

Z-Wave is complicated :man_shrugging: - I’m not really sure what you want to hear or expect here. Writing a Z-Wave library like OpenZWave or zwave-js is not a simple task, and that’s a lot of the reason why up until a few months ago OpenZWave was the only open source Z-Wave library.

Home Assistant doesn’t have the resources to build its own Z-Wave stack from the ground up, zwave-js didn’t exist when the new integration was written, and OpenZwave is the only other open source solution out there.

15 Likes
98

I have a dumb question. Is it the case that HA interacts with qt-openzwave through an MQTT protocol, and otherwise, it doesn’t know what/who it is interacting with? Or is that not really the case.

If so, is this protocol reasonably generic, so that if someone wanted to write a different backend from openzwave, they could implement the MQTT protocol, and it should interact with HA the same way?
Maybe I am missing something - I’m just getting spun up on HA again after two years of freezing an old version because it just worked for me and I was tired of thrashing with every update. FWIW, I tried standing up a new HA instance and using the zwave add-on and had a lot of issues with it. I had a bunch of missing entities and didn’t understand why, for example.

1 Like
99

As a new user of HA, may I throw in my two cents here? I’ve been using Domoticz for the past five years to manage my small Z-Wave network of 30 devices. I’ve become somewhat apprehensive about the project management of Domoticz, so a month ago I looked at HA as a possible alternative.

It was quite clear to me that there are two “official” Z-Wave integrations available, and equally clear that QT OpenZWave is a beta. Nonetheless, that was the one I chose to evaluate because the pros (e,g, support of OpenZwave version 1.6, and running in a separate Docker) outweighed the cons of it being a beta.

To date, the beta has performed well enough in my network to carry on using HA with the beta as the platform for our Home Automation system for the foreseeable future.

3 Likes
100

I get this feeling:

The HA “core team” hides behind the “it’s open source” statement. And advocates to be reliant on “single persons that put spare time in this”.

Although I fully understand that’s commendable (the person), it’s also sad.

In my opinion, open source =! single person.

Zwave is not some little bs protocol. Yes it’s complex, but so is zigbee and esp. Again, if HA wants to beat the competition, put more effort in zwave and make 1 solution stable. You (HA) really have gold in hands! Get organised and structured in you dev process for this.

And I would like an answer from @frenck.

Thanks again for all (besides and indulge zwave support)!

4 Likes
101

That’s what most of us experience. Now it’s time to step further.

For you it’s just waiting for a hickup… and then you’ll understand.

1 Like
102

This is the gist and that’s all there is to it:

There are volunteers available who are working on OZW via MQTT. We’re collaborating and doing our best to make that as great as possible.

If you want to build and maintain a different approach on integrating Z-Wave, please do so. We will support you too.

That’s it. That’s all there is to it. That’s the open source way.

5 Likes