First, I have been using KDE happily for years and considered it far superior to Gnome. All that changed when I tried to upgrade to Fedora 10 which uses KDE4.
A couple of examples.

By default, KDE menu no longer works by displaying a series of nested menus, each beside its parent, with the contents of the submenus being visible as previews before you even descend into them; instead, it presents a kind of sliding keyhole view. This may be visually smooth but it's functionally clunky: it's harder to see where you are in the menu hierarchy and it takes longer to find something if you don't remember which submenu it's under - and even if you know where you're going, it takes more time and more mouse movement and scrolling to get there.
If the keyhole-view menu widget had been the original one, everyone would be delighted at how much of an improvement the ordinary menus-appearing-beside-menus one was.

Yes, yes, it's possible to change this widget back to the way it used to work but that wasn't immediately obvious. And if such clearly dysfunctional eye candy really must be included at all, why make it the default? Aren't the people who would be impressed by it exactly the same people who would enjoy spending time seeking out fancy options and special effects?

Minor but illustrative: KMahjongg is a game I play occasionally when I need to get my brain in gear so it was one of the first things I tried. The tiles used to have a solid design with good use of colour so that they were easily distinguishable. Now the tiles have been changed so it looks a lot like the Gnome equivalent: to the interesting difficulty of finding patterns and planning sequences of moves has been added the merely annoying difficulty of trying to tell the tiles apart.

I could go on. A lot of panel widgets that didn't seem to work at all. Worst, after some tinkering with the size of the KDE panel at the bottom of the screen, most of the screen was lost (except for the panel itself and another thin slice at the top). At the next login, I couldn't even get KDE to start up again: there was just a black screen, with a dark grey rectangle in the middle. The whole KDE environment had crashed permanently.

Perhaps some of the bugs have been fixed. I've also now heard that 4.0 was apparently an alpha release and 4.1 a beta release. If so, it was irresponsible to give them major release numbers without any alpha or beta label.

And then there are still those design mistakes like the ones mentioned above - change for the sake of change. It is really, really disappointing.

Thanks for doing such a great job up to KDE 3.5.
From now on, I'm looking for an alternative.
Michael

michael777 wrote:dysfunctional eye candy really must be included at all, why make it the default? Aren't the people who would be impressed by it exactly the

a. Flawed or not, Kickoff was the product of a study
b. There was no one else working on a menu on the pre-4.0 days so the only step that was taken was to port kickoff
c. There are alternatives now (Lancelot)

I could go on. A lot of panel widgets that didn't seem to work at all.

Please be specific. "didn't seem to work at all" doesn't give any clue on what not might be working.

was apparently an alpha release and 4.1 a beta release. If so, it was irresponsible to give them major release numbers without any alpha or

What version did you try? Besides, this stuff about "4.0 should have been named..." it's a thing of the past. It's been a year since that and the DE has made huge strides (did you take a look at KDE 4.2?).

above - change for the sake of change. It is really, really disappointing.

It's not like the KDE developers sat around a table and dancing like monkeys decided "ok, now what we'll break today...": there are a lot of improvements, both under the hood, and user-visible. No one "changed for the sake of change", if so, why spending 2 years before 4.0 was released?

einar wrote:a. Flawed or not, Kickoff was the product of a study

Really? Can you point me to that study? Because I also find Kickoff to be incredibly annoying. In terms of usability, it takes more mouse manipulation to get to a particular menu item than does a classic menu.

Brandybuck wrote:

einar wrote:a. Flawed or not, Kickoff was the product of a study

Really? Can you point me to that study? Because I also find Kickoff to be incredibly annoying. In terms of usability, it takes more mouse manipulation to get to a particular menu item than does a classic menu.

Use Google, Luke!

http://en.opensuse.org/Kickoff
http://conference2006.kde.org/conference/talks/33.php

Thanks for the link. Some observations:

1) Discounting item 12, task times between Kickoff and classic are roughly the same. Kickoff is faster for some, and classic for others.

2) Item 12 is a huge difference. But look at what it is, adding a document as a menu item. It must have been different in OpenSuse, because in KDE 4.2 the way you do this is IDENTICAL for both menus: use the menu editor.

3) Look at the tasks themselves. Are they common tasks that the user will perform dozens of times a day, or tasks they might perform only once a week or less? It's definitely skewed towards the latter. The classic menu wins with the two most common tasks, locking the screen and shutting down.

4) Most usability studies ignore the expert user, and this is no exception. Where is there any practice time? Where are the groups already familiar with the menus? This study shows that a newbie to KDE4 will find Kickoff to be more usable, but it says little about intermediate or advanced users.

Please note before flaming that I am not criticizing Kickoff! I think it is a very good menu. I find it's search to be far quicker than krunner. But as an expert user, I find that it frequently gets in my way. For example, having to use a scrollbar to access many items.

IMO, an expert user does not need an application launcher. I use kickoff only when I need to add applications to the panel: for the rest, I always use krunner.

Brandybuck wrote:Thanks for the link. Some observations:

1) Discounting item 12, task times between Kickoff and classic are roughly the same. Kickoff is faster for some, and classic for others.

I completely disagree. Kickoff is faster in 9 of the 13 tasks, equal in 1, and only slower in 3. It is over 25 seconds faster in 8 of those 9 tasks, but over 25 seconds slower in only 1. It is over a minute faster in 5 tasks, but again only over a minute slower in 1 task (the same one). The task it is worst at is task 8, but there are three tasks where it beat the older menu by a larger margin.

I think saying "Kickoff is faster for some, and classic for others" gives the impression there is some balance between the number it was faster at vs. the number it was slower at. But it was faster at 3 times as many tasks as it was slower. Even for its slowest task, there was 3 times as many tasks that it was faster at by a larger margin.

It was also rated significantly higher by the users:
http://en.opensuse.org/Kickoff/Results_attrakdiff
http://en.opensuse.org/Kickoff/Results_isonorm

And had a perfect success rate in every task, while the old menu had a 10% failure rate in 3 tasks and a 100% failure rate in one:
http://en.opensuse.org/Kickoff/Results_ ... andmethods

You're missing the point. Two and a half minutes versus three and a half minutes may look significant on paper, but if it's a task you only perform once a month, who cares? In these cases the differences are far less significant. Actions that are performed frequently show classic style menus edging ahead.

I've spent all day using Kickoff, just to give it a fair shot. It is geared towards two use patterns: the guy who typically only runs three or four programs, and puts them in favorites; and the guy who uses the search bar for everything. For someone like me who uses a wide range of applications, Kickoff tends to get in my way.

Different people will have different needs. Is that really such a radical idea?

p.s. I would love that search bar as a standalone applet in my panel.

Turning the computer on and off is a pretty rare task, at least for me. I do it maybe once or twice a week. I can't see someone doing it more than a few times a day. However, tasks like launching an application (6 or 10) or accessing a file I want (2, 3, 4, and 7) are extremely common, being done perhaps a half dozen to a dozen times a day. For every time you turn the computer on, you will launch at least one program, probably several, and probably look at several different documents. So launching programs must be a more common task than shutting down a computer.

In both the application-launching tasks kickoff was better, with task 6 being around 30 seconds better. If you look at the file-launching tasks, kickoff was over two minutes (over 125 seconds) faster at task 2, 25 seconds faster for task 3, maybe 5 seconds faster for task 4, and 25 seconds faster again for task 7. So far the tasks people do most often with an application launcher, that is launch applications, kickoff is considerably better. For the other most common task, accessing files, it is even better still

In fact that only things that the old menu was faster at was removing an entry from the file manager (an extremely rare occurrence) and shutting down the computer, which you probably don't do more than one or twice a day and which I have dedicated buttons on the panel for. Actually, I think having the shutdown/lock button widgets on the panel is the default, so for users who stick with the default configuration they should never have to touch the application launcher to restart or lock the system.

Different people will have different needs. Is that really such a radical idea?

If you had just said that, we wouldn't be having this discussion. If you just have some sort of vague and subjective dislike of the kickoff menu, I guess that is fine although it would be better if you had specific suggestions on how to improve it. But you gave specific reasons for dismissing the usability study, reasons I do not think are fair or valid.

For someone like me who uses a wide range of applications, Kickoff tends to get in my way.

Here is a good example. Exactly what about kickoff gets in your way? What specifically makes launching a program from the menu in kickoff slower than in the traditional menu? If we can figure that out perhaps it can be fixed.