My own opinion is Deskbar needs a built-in hot key “quick launch” menu that scans through all the application menu items and lets you rapidly type them in to search / select / launch.
I know we have 3rd party tools which do something similar, but it really should be built-in behavior since all the modern desktop UI’s have gravitated to it.
I don’t quite understand why you’d want this build into deskbar? We already have some third party tools doing this right now. Adding one to the image and binding it to cmd+space like macos would be quite doable. Deskbar should stay a “slow” launcher where you can take your time to look through the options, albeit with a redesign like moving the menu items one layer up like proposed.
We need the main way to launch apps to be slow? Now the arguments are getting weird…
Deskbar has its limitations. This was already discussed many times, and the solutions already agreed on:
The current DeskBar menu is worst of all options: tiny menus that are hard to reach, tiny icons that are hard to see, no categories, no way to filter down the list, and a single colon so a long list still won’t fit in the screen despite everything being small. It’s time to admit the best practises in that area have moved on and there are better ideas out there than BeOS had 25 years ago.
That’s a rather selective quote of my post, no? I said in the same breath that quicklaunch (if that is the one where you type) or equivalent should be included aswell. Those are simply two different usecases/ways to launch stuff and trying to smash them together does not work well in my opinion.
You can replace the word slow with “mouse focused” if you want. The current way of deskbar works nice for the case of having a list with all installed stuff, that part is good. Just the menu above that needs to be redesigned, as proposed here. I don’t think it needs to be like quicklaunch, because you should just use quicklaunch instead.
(MacOS also has this distinction between the cmd-space keybord view and the app list, just that it is a fullscreen grid instead of a list)
Fully navigatable with keyboard!
I love it!
Great work!
I think it isn’t.
Making mouse and keyboard be exclusive to each other seems like a bad idea. Keyboard to quickly filter, and then mouse to select from a more limited set is likely the most efficient way to do things. And I think in this case, it can be done without much downgrade to the “all-mouse” approach.
I still do not get what we are trying to fix, by just replacing the “leaf” button with four new ones. That will make the Deskbar bigger for no purpose (just to save a click when launching an application?).
just changed a little the layout so strips contains more icons, it looks better now, tnx for the hint
you compared this interface to a clone of the Windows 11 start menu.
It might be interesting to consider that the use of a filter field and icons in a squarish window represents an established paradigm for a app menu, present in many desktop interfaces like those of KDE, MATE, and Cinnamon…
Regarding the icon strip, I understand that it’s not perfect, but it’s a trial implementation that, although not visible in the video, allows drag and drop* to the Favorites strip (Be Menu folder) and to the Go (Go folder) - the Go folder is somewhat hidden and currently accessible only through the file panel, but it’s useful for quickly defining locations for opening and saving objects, I think it deservers to highlight this feature.
*As for the BMenu, it seems that it does not support drag and drop or contextual menu opening
This implementation could address the current Deskbar limitation, allowing for the organization of the BeMenu without having to open its folder and create symbolic links.
“It looks like some Windows 11 clone that is not good idea.”
+1
I do agree that the deskbar needs some love but the last suggestions mostly look totally crazy to me. The thing basically works fine and just needs some refinement, not a total redesign.
I think this is a good enough mockup that having icons there is pretty much useless - they all look the same. Icons may only work for applications, but even then, it should always show the text, too.
In any case, I think the usability of Deskbar could be greatly improved without creating a completely different experience.
I fully support this, how can i start QuickLaunch with SUPER Key ?
I only aprreciate the search part the rest is too much…
Applying a label would not be a problem 
We already went through an evolutionary , and revolutionary, change with packagefs that disrupted part of the user experience; some still complain, but that change was necessary and I won’t return to the old installation methods.
I realize that UI/UX is different, the current deskbar has been a long-standing habit for almost 25+ years, so suggesting a different approach can seem unusual or even, as someone wrote here, “crazy” — dang man, that feels a little harsh !
The current deskbar fails according to Fitts’s and Hick’s laws in several ways: it has a long vertical menu with nested menus (so adding categories to application list would be worst), small elements, and seldom-used items at the top.
Probably I’m wrong because I’m not a programmer , but I think it’s difficult to implement a search in the current menu because it doesn’t use a BView where you could add a text control and a BList (like QuickLaunch does); instead it uses a BMenu with BMenuItem, maybe it may be subclassed ? but the previous problems of decision noise for the user would remain, so if the deskbar has to be redesigned — at that point it’s better to rework it to be more usable.
My design, while far from a perfect solution, better follows those principles: larger, more easily clickable elements; visually separated sections; menus shown only on user request (right-click); it features drag-and-drop completely absent on the current deskbar; quick text search (the text control would receive focus when the window is opened); and shorter action times compared with the classic deskbar, a group of talented programmers as you all could find a much better solution.
What I’ve shown here reflects my own needs and may differ from those of other users.
The mockup is intended to start a discussion about user needs that the existing deskbar does not currently meet, but instead I’ve perceived a little too much rage, I’m sorry if this hurt someone’s feelings, but it was unintentional.
Obviously not every need can be satisfied, and compromises will be required.
The paradigms in this mockup are similar to features already used in other programs, such as LaunchBox for the favorites/go icon strip, and the type-ahead filter in Tracker or QuickLaunch for the app search, so a similar approach won’t mislead an experienced user (and probably neither a rookie).
actually you can using “shortcuts” , define your keybinding to /boot/system/apps/QuickLaunch
The package changes were not really the same, that was a change from having no packaging to having full system packaging. The lack of package management was a glaring hole in the system, a problem that was you could argue was stopping the entire OS from progressing.
UI design is different… it’s far more subjective than something like packaging. We have one data point at present: The deskbar works quite well for most people but there are some common pain points. So something similar that addresses those points should also work quite well. It would also be in keeping with the current system and BeOS. On the other hand, a total overhaul is a total unknown.
If we look at the complaints we have about the deskbar, most users are experiencing niggles, not major problems: There should be categories, the entries should be in a different order, it would be nice if there was a search/filter box. They dont have any other issues that would warrant a total redesign. That is why it seems crazy to me to completely change the thing into something totally different: Fixing the issues that have been reported just doesn’t require something completely different to what is there already. And I dont know if you consider the proposed changes a refinement, but personally I struggle to recognise the current deskbar in those designs.
We can look at analogy: Someone is designing a car, and testers have a list of issues they found with it. The solution is not a different car (with a totally different set of issues), it’s to refine and improve what is there already.
I dont want to discourage innovation. As a research project it might be interesting to get more feedback about a working implementation of new ideas. Could it be a different deskbar that one could install and try out, or an alternative mode? But I think suddenly changing to a totally different design is a mistake.
It looks good, but there’s definitely room for improvement. The icons are not intuitive. Additionally, you’ve suggested a dark color design, but if this is indeed a discussion mockup, a good starting point would be to use the default theme and colors - light gray.
I think your design is too far off deskbar to be a replacement; however there is no reason why this can’t be implemented. Deskbar can be stopped with the launch server and something else be run in it’s place.
The only problem is twitcher beeind in deskbars team, but this should be splittable.
Why can’t you have both accessible by the user? I could see maybe some hot key that could fire this what I would call Deskdrawer and still have Deskbar hanging around. Of course there could be a preference for offering one or the other or both. Twitcher needs to stick around regardless.
It’s not needed to put everything in the default install, but on the other hand there is nothing stopping you from installing something like this from the depot.
