Hi @StephaneB,
I understand your message!
I’d like to bring some additional data to help you better understand.
Already, @cicoub13 said it: the features you see in « Demandes de fonctionnalités » are the ones that remain, but you don’t see all the ones that were developed and are in the category « [Archive] Demande de fonctionnalités - Gladys Assistant community ».
A quick query on the forum database gives me:
So 138 archived, for 258 current requests!
So before reaching the top features you see now, there were many other requests that were implemented, and which may seem obvious today, but didn’t always exist.
A few examples: Live camera stream on the dashboard, addition of an « Devices » box not dependent on the room, Sonos integrations, Zigbee2mqtt integration, Google Home integration, etc… (There are 138 of them!)
On a case-by-case basis
I agree with @cicoub13’s list:
- Continue only if … ELSE / ELSE IF: Two ways to see this request: either we make a « hacky » solution calling another scene in the else branch, except that doesn’t really solve the problem. Or if we want Node-RED-like scenes, we need to totally rethink scenes, which represents months of work. For me there is no consensus nor an easy solution on this topic.
- Dark mode in Gladys: As @cicoub13 said, we depend on the update of our Tabler theme, which is slowly moving toward 1.0 but remains in beta for now. We’re waiting for that 1.0 because the work to migrate all components will be titanic and we don’t want to do it twice.
- Add non-detection duration in scenes: Vague request, we don’t know what needs to be done. After setting up my smart home, I realize that this information is usually stored in motion sensors, so I even question the usefulness of this request.
I took the top 3, but I’m willing to discuss any request you have — feel free to re-open the relevant topic if you have specific requests.
Voting isn’t everything
I don’t know if you’re familiar with IT project management, but there are prioritization methods because this topic is recurrent in every project 
Whether you’re a small open-source project or Apple, there are always more requests than developers, so prioritization is necessary.
Several methods exist:
(Source: Prioriser les fonctionnalités Guide des frameworks & roadmaps | Arquen)
As you can see, the « vote » is only one part of prioritization (the « Reach » in the RICE method), but it is weighed against other factors, notably cost/effort, which is just as important.
A feature that is highly requested but takes 1 year of work for 2 people is certainly less attractive than a highly requested feature that takes 2 hours to develop.
Personally, I’m not tied to any particular method, but this is the kind of thinking that goes through my head when I decide what to work on:
What will have impact and at what cost?
Another important point is that votes only represent the viewpoint of the existing community (a few hundred active members on the forum).
However, if we want to attract more users, some requests are necessary but will naturally be absent from the requests because the targeted user isn’t on the forum right now: it’s the chicken-and-egg problem!
So keep in mind that forum requests are only part of the spectrum: we must keep our ears open outside the community.
This reflection led me to prioritize 2 tasks recently: External MQTT broker on the Zigbee2mqtt integration & custom MQTT topic per device.
Conclusion
So, as you understand, it’s a complex and always frustrating topic because we can’t develop everything. The important thing is that the project moves forward
If you think of particular requests, don’t hesitate to ask me what I think — I’ll be happy to answer.
Note: I’m expressing my personal point of view here, but this does not commit any contributor, as I remind you everyone is a volunteer and does this in their free time. Each contributor is free to develop what they like ^^
