Compared to v3, the nice features in my opinion:
- Remote control is complete control of Gladys (it’s the same UI as local Gladys, whereas in v3, due to technical limitations of v3 it was a different, very limited frontend)
- Automatic backups every 24 hours
- The open API
- And indeed the new integrations that are coming (Google Home, Alexa, etc..)
No problem, I wasn’t sure about the intention of the post, and if you ask the question you’re not the only one to think that, so I took the opportunity to give a complete answer 
On this point, I think you are confusing several things:
- On the one hand, Gladys Assistant is an open-source project under Apache 2.0 license. Anyone can develop an integration on a voluntary basis. I do my best to motivate the community, I answer on GitHub, on the forum, I regularly have one-on-one calls with anyone who wants if someone is stuck, I try to be the « orchestrator » of this community.
However, these contributors I cannot force to develop, everyone has different life projects, children arrive, a job, and Gladys is often a side-project that comes after all that in the evenings and weekends.
In general, developers on the forum only create integrations for the hardware they have themselves. One should not believe that these devs are charitable associations, they develop primarily for their personal interest, and if they have the time they share their work with the community.
- Gladys Plus is a company, and I have a responsibility towards Gladys Plus customers. I respect a certain quality of service, and I listen to their requests to keep this customer base.
Now to get back to the integrations and my personal time allocation.
In v4 it’s like in v3, I personally try to invest my time and expertise in what could have the most impact on the project: the core of Gladys.
For example, this summer, I worked on different subjects:
- The multi-dashboard (a rather big piece)
- The aggregation of sensor values on curves on the dashboard (a rather titanic work that took up a good part of the summer, much more than the Google Home integration in terms of development time)
For me, these are fundamental subjects that are very important, and if I don’t get to them, no one will work on them. These are subjects that are complex in terms of UI, and complex in terms of data processing.
As these developments are measured in weeks of development, knowing that I am part-time on Gladys, so the development is spread over several months.
Because developing an integration is not a one-time development, it’s continuous maintenance.
When an integration is developed by an external dev, usually the dev takes care of the maintenance of the service and is a bit the « referent » of the service. It’s a model that works, and the integration is stable and sustainable over time because there is someone who knows the code and who uses the integration on a daily basis.
Example:
- Caldav by @bertrandda
- Philips Hue maintained by @VonOx
- Bluetooth by @AlexTrovato
- Zigbee2mqtt by @cicoub13 & @VonOx
If I had developed these integrations myself, I would be the only one to know the code, and I would have to maintain the service (add new devices, debug the different problems that arise/do updates).
Given that if I have to take care of:
- The community and all the communication around it (YouTube videos, blog posts, site maintenance, etc..)
- Core development
- Management & development around Gladys Plus
All this on a part-time basis (2.5 days/week), I realistically don’t have the time to get into these integrations.
Finally, I think it’s always better when it’s a dev who really uses the peripheral who does an integration, because he understands the need better.
For example, I went several days to @hamtaro’s in Niort to use all the Xiaomi gear he has, and develop the first Xiaomi service in Gladys v4 in order to « initiate » the development of v4 service and show a first service to the community.
The problem is that I was only a few days at his place, and as I coded the service, there was no « referent » (in theory, it’s me the referent: except that I don’t have the devices!)
So since then, I haven’t been able to touch the integration.. I did look on the forum for a replacement, I didn’t find anyone to maintain the integration (Fortunately since then there is Zigbee2mqtt).
To conclude, for me the approach « one integration = one referent who really uses the gear at home » is the right one, and I think I have more impact on long-term subjects like the core.
Top, it’s coming in the coming weeks
Why not, I sent an email last week to a selection of former contributors, and otherwise I’ll talk about it in the newsletter I send (about 3000 subscribers).
Edit: Regarding the lack of DEV on the Z-Wave integration, we were talking about it the other day here: Z-Wave - Gladys v4 - Windows WSL - #121 par pierre-gilles