As suggested by Pierre-Gilles, here’s my list of small UI improvements that would make the user experience in Gladys more comfortable. If I understood the idea correctly, everyone can comment on this list (to agree, add to it, or contradict it), and every idea that reaches consensus will become a Feature Request.
@pierre-gilles: If others have other ideas, should they make another post (in which case I’ll add my name in the title of that one), or should they continue on this post?
when a column already contains several modules and you want to add one in the second position, it’s a shame to see it get added at the very bottom and then see it move up afterward
when I add a chart, I almost always choose the same settings: a line chart, with axes, without the variation,… It would be better if these settings were set by default. Either by defining them in my profile, or by reusing the settings of the last chart added.
In a « scenes » module, I’d like to be able to define the display order if I show several, rather than the alphabetical order
when I add a device (in a ‹ chart › module or in a ‹ devices › module) and there are many, I sometimes have to fumble because I no longer remember the exact name of a module. It would be convenient to be able to filter the dropdown by room, or by device type…
you can’t simply move an entire column. For example, move the right-hand one to put it in the center.
In dashboards, in view mode
On mobile, there are still cases where values are not updated automatically and would deserve a manual refresh: this is the case, for example, when you leave Gladys for a few minutes to use other apps and come back: the values or graphs are not up to date, and will only be updated if a new value is received. A ‹ refresh › button would be really useful (or support for « pull down to refresh »)
In the charts, the time scale is not homogeneous enough; there is, for example, a big gap between 1h and 24h. « 1h, 24h, 7d, 30d, 3m, 1y » gives gaps of x24 x7 x4 x3 x4. At minimum, we should add a « 4h » or « 5h », I think. And maybe also a « 3d ». Or even change more values for a quasi-regular scale (all x3 or x4): « 30min, 2h, 6h, 24h, 3d, 10d, 30d, 3m, 1y »
In a chart, the box that appears when you hover the graph with the mouse is too often placed above the cursor, so you can no longer see where the cursor is
when you type letters in the search box, the search is triggered immediately on every character, which means that if I type the letters « at a certain speed » (neither fast nor slow), the search runs after the first letter is typed and erases the second character when the result arrives… I don’t know if I’m being clear
sometimes, when filtering by a tag, the filtering works but is only shown for an instant, then all scenes are finally displayed (even though the desired tag stayed checked)
sometimes I want to find all scenes that use a certain device. For example to modify a scene before deleting a device. It would be convenient to be able to filter scenes by that criterion.
In scene editing
when a scene becomes complex, I would like to be able to comment it to remind myself of its logic. And if I put this kind of info in the description, it is displayed on the page listing scenes. There should be a ‹ comment › field in addition to the ‹ description › field (or even better, be able to add a comment to each step of the scene)
when duplicating a scene, its title is empty even though duplication is often intended to create a variant of the scene, so with a title that could be similar. That forces copying the title to paste it afterward; it would be more convenient if the title were pre-filled with a « copy of [original scene title] »
when you want to add a step in the middle of a scene, you have to add it at the bottom then move it up. It would be more convenient to be able to insert a step.
when a scene is long, finding a particular step is not very practical. It would be nice to have on the left or right of the scene a ‹ menu › showing the different steps and be able to click on the step to which the view should jump. And it would be useful to optionally give a name to each step so it’s more readable than a simple list of step numbers
when you have several « retrieve last state » actions and use them in other steps, deleting one of the « retrieve last state » actions sometimes causes a shift in the numbering of the ‹ retrieved variables ›, and you have to go through all the steps to adjust… the renumbering should ‹ propagate › by itself…
I know there’s already a feature request, but I’ll mention this need anyway: the absence of « if… then… else… » forces duplicating many scenes, making maintenance of common sections more complicated…
when I add a device, all its entities are added with history enabled. And I don’t always remember to go and adjust that even though I don’t need to record everything. It would be handy if adding a device opened a window asking me to confirm which entities I want to record in history or not
the « allow pairing » button is not properly synchronized with what I see in the Zigbee interface
@pierre-gilles there haven’t been any particular reactions to the UX/UI improvement ideas I listed here (only @prohand, who appreciated the first two batches). Do you consider this a consensus, and should I therefore create a feature request for each idea?
I was waiting for some feedback before making my own so as not to monopolize the discussion, but if no one jumps in, I’ll go ahead
Dashboard
Could this be solved by an easier drag & drop that would « compress » the size of widgets, like we do on mobile for example?
I see two things:
First, we could set default settings instead of leaving quite a few parameters « empty »
Indeed, we could save the last settings of the last chart, that doesn’t bother me!
That’s totally possible!
On one hand I agree, but on the other I struggle to see how we can fit that in such little space. If we add a search bar + a filter-by-device-type bar, that adds quite a bit of height
But it’s worth investigating, I agree with the comment!
Why not!
I think the opposite: it should be automatic, indeed currently we don’t handle refresh when you come back to the application, and we could handle it fairly easily. I vote for it!
Why not, but then we lose the last hour?
I’m really not against changing it but we need to gather the needs of other Gladys users.
Okay to start the debate on a specific topic.
Why not change that indeed (as far as possible, we use ApexChart, I don’t know if it’s configurable)
Scenes
You’re not the first to complain, I created a GitHub issue, it’s a bug:
The problem with that is that the title then becomes the « selector » (unique identifier of the scene), and so you’ll end up with scenes called « copie-de-detection-mouvement-entree », instead of the proper name of your scene. Knowing that the selector is immutable.
Afterwards, it’s a choice, we can say we don’t care about the selector, but then it creates little points of frustration, for example if you use the Gladys API to trigger scenes, then to identify your scene you’ll have this title which doesn’t correspond to the current name of your scene.
Now, which point of frustration is stronger?
Why not!
As a result, we lose screen width on the scene view which is often already compressed! To be seen, I’d welcome a mockup
Other scenario creation software can be a source of inspiration (eg: Zapier, Apple Shortcuts)
I’m afraid that would complicate the scene creation interface a lot, which in my opinion is already complicated…
I’d like you to check if this is something that’s possible in other scenario-creation softwares, and if so how they do it (Apple Shortcuts, Zapier)
Yes! There’s already a feature request for that I believe (I think it was @lmilcent who created it but I don’t remember). I’d like you to find it and upvote
I know, and I’m the first to find this limitation inconvenient ^^
For now, no one has found a solution.
I’d be keen to know how other scenario-creation softwares do it (Apple Shortcuts, Zapier)
Zigbee2Mqtt
I understand the problem, but I’m not a fan of that solution.
Maybe that view could be simplified so that logging is more visible at a glance?
Indeed, I don’t think the synchronization is done both ways / nor that the current implementation supports the new pairing mode with expiration that appeared in Zigbee2mqtt after Gladys integration
Thanks for your review. Should I file change requests for everything, or only for some items? Or should I wait a bit longer for other opinions? Or do you handle things as they come, like you did for the ‹ search lag ›?
Tell me what is most convenient/effective for you…
I’ve already experienced this behavior, but it’s not every time, so it’s hard to reproduce in my opinion.
On the software side, it was with Firefox 132 to 134 on macOS 11. I moved to macOS 13 — I’ll see if it continues and make a video if I can, but generally it only happens once; afterwards the tag is applied correctly by unchecking and rechecking the tag.
I just did quite a few tests to reproduce it while I was recording. I isolated the interesting part. In the video, I only make 4 clicks: one on the filter button, one to check the ‹ bus › tag, one to uncheck it, then one to check the ‹ water circulator › tag.