The Linux Box has interviewed Aaron Seigo on their latest episode of The Linux Box Show. He discusses Appeal and the plans for making KDE 4 the leader for usability, development and cool eye candy. Specific topics he covers include KControl, package management, KOffice and using high level programming languages. Start 5 minutes in for a brief history of KDE and 10 minutes in for the interview, or read the transcript.
KDE 4 and beyond: The Linux Box interviews Aaron Seigo
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
The interview in question.
http://aseigo.bddf.ca/?pID=1221
Hope KDE 4 delivers on this. Improved useability is its only real barrier at the moment.
well i find it nice and useable allready but then im no 70+ grandma…
Reducing clutter in toolbars is good, but don’t forget about menus. Just look at the disaster that Konqueror’s Settings menu is (or what is it called — don’t know for sure, because I’m using localized version).
well, imho the settings menu isn’t that bad, but they should clean the window menu – there are nice things there but no-one uses them… the can be accessed by the tab bar (rmb), and that’s much easier.
One thing they should focus on is remote administration.
If they haven’t checked it out allready it’s high time they see how Windows2003 and Novell system works, it’s really great.
The kiosk project is a start, but it’s still far far behind MS Windows(and the system Novell sell I guess, but I haven’t used it).
Reducing clutter in toolbars is good, but don’t forget about menus. Just look at the disaster that Konqueror’s Settings menu is (or what is it called — don’t know for sure, because I’m using localized version).
I second that. Here it is in all its glory:
Hide Menubar
Toolbars ->
———————————–
Full Screen Mode
———————————–
View Properties Saved in Folder
Remove Folder Properties
———————————–
Load View Profile ->
Save View Profile “File Management”
Configure View Profiles …
———————————–
Configure Extensions …
Configure Spell Checking …
Configure Shortcuts …
Configure Toolbars …
Configure Konqueror …
There’s so much to criticise here, e.g.:
– Why is one of the items called “Configure Konqueror…”? Isn’t that what the whole Settings menu is for?
– Prefixing half the items with “Configure” is a distracting waste of space. The whole menu is already called “Settings”, so just “Extensions …” would do fine.
– “Toolbars ->” vs. “Configure Toolbars …”
– “View Properties Saved in Folder” is too long and still doesn’t really say what it actually does. To me the phrase indicates a list of properties or something popping up. In fact it’s a toggle option with no immediately obvious consequences.
These are ambitious goals indeed!
Appeal seams to be something I’d like to contribute to. Maybe I’ll find the time to do more than just lurking. With all the new exciting X.org improvements to their X server, this seems like to be something where tools to make the Appeal goals happen are going to be available soon.
The new X.org stuff offers a lot more potential than just “eyecandy for the eyecandy’s” sake.
What Seigo hints at is really a huge field of work, but probably a very rewarding one.
I would like real SVG icons, not pre-resized icons. I know it would eat some CPU cycles, but it would make themes wonderful in all resolutions, and icon resizing when a mouseover occurs would be great!
that means i’m looking at April 2006 before i can get a KDE4 based SUSE release.
>One thing they should focus on is remote administration.
You are not very well informed.
KDE already has the best remote administration system (kiosk) available anywhere. It is the best system because the system is actually *used* by the KDE apps. You can lock down and configure remotely almost everything about all KDE apps.
If you are talking about system wide configure tools (not desktop wide), it is out of scope for KDE. In this case, KDE should offer a system that is easily plugable, for any system like this easily control the desktop as well. And it is already done. (The reason is: system wide admin tools are OS specific. KDE runs in many different OS.)
A curiosity: how could you make such a claim (that KDE lacks good remote admin system) without any information at all?
Most KDE SVG developers stopped to work on KSVG due to some limitations that they found on their approach. So they started over.
Now KSVG2 is getting there, with the ability to cover much more of the spec. SVG can be a flash replacement, and KSVG2 already supports animations.
KSVG2 is planned to ship on KDE 4.
BTW: on KDE 3.4, there is the ability to use SVG wallpapers
The interview is a great insight into the efforts of the APPEAL project. APPEAL works, in Aaron’s words, to bring “Breathtaking Beauty” combined with “Clarity in the UI” to future KDE versions.
In addition, new innovatie concepts are being designed, to manage the zillions of files, settings and objects populating modern computers:
–> a “context linking engine” named Tenor will be complementing the more simple indexing searchtools
–> a “content browser” going far beyond the traditional file manager idea
–> a search driven interface to configurations instead of the hierarchy of dialogs and menus
As Aaron says, back many years ago, when GUIs for configurating systems, programs, menus, toolbar first appeared, there where only a few dozen items to deal with. A “hierarchical” organisation made sense then, users could easily learn and remember these.
Now we virtually have hundreds or even thousands of user-accessible settings: for system, programs, menus, toolbars, thems…. This is true for *all* OS platforms, not only Linux. Users feel lost in that jungle. I can see it every day with collegues using Win XP Pro and Win 2000. (KDE is more configurable and flexible than everyone else — so the “itch to scratch” is more acutely felt here).
The same is true for the files on computers. We used to have a few hundred in DOS, or a few thousand at the most. Now we have up to a million or more on many users’ workstations. They are organized on disk into a hierarchical structure of directories, paths and filenames. Traditional file managers are designed to work with that hierarchy. It was easy to navigate when there were not as many files and directories as there are now. File system navigation with file managers now it becomes more and more painful.
At the same time we are not putting to disk much of the info that our brain desperately tries to recall when we search f.e. a photo. “Wait… it was sent to me as a mail attachment from Sylvi. Hmmm… where did I store it? Or do I still have that mail? It came around X-mas….”
The idea is to let a “context linking engine” do the memorizing work. Tenor helps to collect all this meta and context information, in the background, automatically, and organizes it in a meaningful way.
Applications can utilize Tenor for their own purposes. They also feed to Tenor all meta info related to all their actions. This goes much beyond creating fulltext search and indexing. It is more like our brain works. And search results based on that additional “linkage info” will be more meaningful to the user. (It is probly similar to Google’s page rank idea….).
As Scott Wheeler, the principal brain behind Tenor puts it: “Why is it that Google can find some matching content on the Web faster than we can find a file on our own harddisk?”
KDE’s tighly integration of all kinds of applications will be an ideal foundation to give this idea the prospect of succeeding. ISVs who chose KDE as their target will likely also benefit a lot from that ecosystem. Programmers who target environments with no or much less integration will have to do much more work to come anywhere close.
The “content browser” would be a very meaningful addition (not a replacement) to Konqueror (which is shining when it comes to fullfil the role of a file manager), drawing lots of power from the Tenor, the
context linking engine.
The same holds true with a “Control Center of the next generation”: a search driven user interface may make it much more easy to navigate and use than the current organisation, which still reflects the tree/hiearchy paradigm.
Well, kde 3.4 is no longer ‘cluttered’ imho. Not that it was bothering me too much before, but I think with 3.4 kde has become the most polished Open Source desktop I have ever seen.
If you haven’t tried google mail with konqi recently, you should! It works so much better than in firefox – for instance, you can middle-click to open links (inbox, letters, etc.) in a new tab that still doesn’t work in firefox. Not to mention that konqi has become blazingly fast – in my experience, it is 60% faster than an untweaked (but compiled with optimizations) firefox, and some 20-40 if I tweak firefox (enable pipelining for instance).
When do you get tired of beating the ‘oh its so cluttered i won’t use it’ drum? Really, it is getting old and no longer true. Of course there will be someone somewhere who will find a button or an option in a menu he or she does not need, than will assume that nobody needs it, and will declare that teh clutter is so awful that it prevents him of using konqi or KDE
))) Now look at this: ftp://hatvani.unideb.hu/pub/personal/screenshots/konq34a.png Tell me with a straight face which of the buttons stands in your way of using konqi for filemanagement or browsing?
The problem with all the configure buttons is that they all launch different dialog boxes that configure different things. What they should do is make each dialog box in to a KPart and have one option to configure the application (probably rename the option as well) then load all of those KParts in to one dialog box with and beable to switch through all the KParts from a box on the left side kinda like this: http://shots.osdir.com/slideshows/slideshow.php?release=248&slide=2…
Have each of the icons be one of the different config options.
Im a GNOME user but I like the path KDE is taking, I’ve tried the Kubuntu live CD and I was amazed of the speed KDE has improved, I’d only like to see new the defaults fonts, I think they are to big and the menus look huge, nothing I can’t change.
I really hope they deliver on the breathtaking beauty. I freaking hate the toy/fisherprice look of KDE with a passion. It looks cheap to me. I want something that is elegant and pleasurable to look at. Something that looks fancy but not too busy, and not as industrial as Gnome. I don’t care what anyone says, I think the crystal icons are just gay…..yes, I know you can change them, but still.
Linux needs a UI as refined as say OSX Panther+. I know thats something Novell is working on for next year with NLD10, but thats a long time to wait.
If you haven’t tried google mail with konqi recently, you should!
It only improved because Google recently added a “basic HTML” view. The standard view, which relies heavily on Javascript, still doesn’t work in 3.4. Therefore, some functionality, e.g. “Settings”, isn’t available.
More details here:
http://gmail.google.com/support/bin/answer.py?ctx=gmail&answer=1504…
Now look at this: ftp://hatvani.unideb.hu/pub/personal/screenshots/konq34a.png Tell me with a straight face which of the buttons stands in your way of using konqi for filemanagement or browsing?
You could defend the worst of interfaces with that kind of reasoning.
But as you ask, here’s my list of what could be removed:
– Home: by default it just takes you to your home directory, the Home panel button already does that
– Print: Used only seldomly, the menu option in the standard place is enough
– Find: It’s in the menu. For speed use Strg+F, as you have to go to the keyboard to type in the keyword anyway.
– Lock: What does it do?
– Gears: Unnecessary distraction.
– Clear Location, Go: Who ever uses those?
– “Location:”: Waste of space. Even newbies need be told only once where to enter those web addresses. A tooltip would do fine.
– Search Bar: The location bar can do that.
– Favourites Toolbar: Is one mouseclick (on the Bookmarks menu) really worth all that visual clutter?
Remember, users can always add anything that they do actually use frequently.
Wow. Stuff like search-based filing doesn’t really interest me, but this link engine is an idea I’ve never heard anyone mention before. I think I could grow to depend on it. To summarize the article, lets say you’re doing a research paper on pandas. You download a pic of a panda from ilovepandas.com to your desktop for later. Later, you embed the pic in a document. With the link engine, you’d be able to tell, “oh, this pic I used in my presentation came from my desktop, and it got THERE from ilovepandas.com.” This is the first radical change in filing that sounds indispensable to me.
KDE does already support SVG icons.
I can’t be specific, but I’ll try do describe my reaction to that image.
First, my eyes start to hurt (probably something wrong with my monitor).
Second, I’m overwhelmed as I’m trying to digest the whole image, slightly desorientend, and frankly my eyes just won’t fins a peacful place to rest upon.
Third, with severe diciplin I force my eyes down to that gigantic blue area trying to convince them that it actually contains something interesting.
Forth, after a few more moments of desorientation, agony and pain I finally find the little house with the text “Home Folder”.
IMHO this is just to much work to use a GUI. I’m not saying that I should be the one to measure this by, but this is the reason why I’ve uninstalled all my KDE’s after 5 minutes of testing, I havn’t tried 3.4 yet but I’m getting rather familiar with the pattern.
Let me start saying that Konqueror needs a major review, not because of the reasons you say (klutter), but it needs to be more “task oriented”. Today it is a mix of a file manager and web brower, and you can’t make everybody happy, it is impossible. The solution is:
1) Create a simple browser profile, changing not only the toolbars, but the menus and the configuration.
2) Create a simple filebrowser profile.
3) *Keep* a prowerfull browser profile, for people like me who like to browse the web and manage files in the same window, using the tabs. But review the config dialogs anyway, they are a mess.
With KDE technology, it is simple to create these profiles.
That said, I think it the “remove everything approach” is riodiculous, not based in any usability research, and sometimes it decreases usability, especially for newbies. User interface should be based in research with users, not osnews or slashdot comments.
So here are my comments:
> – Home: by default it just takes you to your home directory, the Home panel button already does that
This is a keeper for file manager, but must have a different icon for the simple webbrowser (home folder) and the file manager (home page). The Home panel performs a different funcion: it browses the home folder hierarchy, and should not be shown in the simple webbrowser profile by default.
> – Print: Used only seldomly, the menu option in the standard place is enough
Not true: an usability study was done, and (surprise) the print button is used relatively frequently.
> – Find: It’s in the menu. For speed use Strg+F, as you have to go to the keyboard to type in the keyword anyway.
Agreed.
> – Lock: What does it do?
It shows the security level of the page. Wasn’t removed for (obscure) security reasons, and will be removed next release, when these issues are fixed.
> – Gears: Unnecessary distraction.
Doh. It gives visual feedback if the page is being fetched. You have one of these with each browser, and for good reason.
> – Clear Location, Go: Who ever uses those?
I do. It is neat.
> – “Location:”: Waste of space. Even newbies need be told only once where to enter those web addresses. A tooltip would do fine.
Agreed.
> – Search Bar: The location bar can do that.
the location bat does that. But nobody knows that, hence the search bar.
> – Favourites Toolbar: Is one mouseclick (on the Bookmarks menu) really worth all that visual clutter?
Agreed.
> Remember, users can always add anything that they do actually use frequently.
It goes both ways. Newbies sometimes have problems finding stuff, therefore, you put frequently used actions in the toolbar. It is easier for an advanced user to remove an action he does not use, than a newbie to add an action it uses a lot. In the end, a usability study should determine that.
“Favourites Toolbar: Is one mouseclick (on the Bookmarks menu) really worth all that visual clutter?”
Well, that toolbar it’s not there by default… And I found the clear location button very useful sometimes.
And I found the clear location button very useful sometimes
That’s the problem, options you just use “sometimes” do not deserve to be there by default.
I actually find that one very useful, much more useful than a bookmarks menu. All my often-visited sites are organized on the toolbar (some with menus, some just plainly there), and the stuff I visit less often is in the bookmarks menu.
Just like the example in the interview half an hour ago I had to find an image for a small web project. I didn’t remember where I had saved it or the name of it, but I remembered well where I got this image from. This new search would have been handy. It’s an use case I have quite often. Very interesting.
One thing they should focus on is remote administration.
If they haven’t checked it out allready it’s high time they see how Windows2003 and Novell system works, it’s really great.
The kiosk project is a start, but it’s still far far behind MS Windows(and the system Novell sell I guess, but I haven’t used it).
Perhaps you should try to use KDE before you critisize it. As a long time KDE user I can tell you there is many things that could be a lot better in KDE , but its network capabilities is certainly not one of them.
In fact this is the reason I keep comming back to KDE everytime I try other systems for a while. When you run Windows or Gnome you sit in front of a computer, when you run KDE you sit in front of the Internet.
The main problem, as I see it, with KDE today is that they don’t seam to know who they develop for. Some things are for perfect the newbie, other are perfect for the advanced user but totally wrong for the newbie. The end result is that it doesn’t work very well for any user.
So, if they succed in their plan to make KDE the leader in usability, they will have no problem getting the users attention. They allready have the technical edge.
Well, kde 3.4 is no longer ‘cluttered’ imho. Not that it was bothering me too much before, but I think with 3.4 kde has become the most polished Open Source desktop I have ever seen.
I agree, 3.4 is much less cluttered. But there is still a lot to do in other areas.
For one thing look at toolbars in applications and the kicker applets. To a user adding a button to the kicker is a very similar action to adding some button to an application toolbar so it would be nice the userinterface for doing it was similar. Preferably it could be some kind of drag&drop system like in Firefox.
Why is the behavior of the trash configured in konqueror, most people would probably think that the file manager and the trash are separate things. It would probably be better if trash was configured from the trash RMB context menu.
Speaking of that menu, why does it have a “Copy” item? Most people put things in the trash because they don’t want it. Making a copy of it makes no sense.
Why is the default position of the trash not in the kicker instead of in the desktop where it often is covered by other windows. Both Gnome and MacOS have moved the trash away from the desktop and from eraly screenshots of Longhorn Microsoft is moving it away from the desktop as well. Why not follow that trend.
Look at the K menu. It is divided into separate parts for settings, applications, and dynamicall things like frequently used applications. The problem is that the menu renders the same regardless if kicker is on top of the screen or at the bottom. If the kicker is in the standard position at the bottom of the screen, you will have to move past all other menu entries to get to the most frequenly used application. This doesn’t make sense. Both Gnome and windows XP is better at this. If the menu was reordeded depending on if the k-menu was on top or the bottom of the screen, it would be possible to place the most frequently used menu item closer to the base of the menu.
Look at the drag&drop behavior of the file manager. When you drop a file over the drop target, a menu pops up that allows you to chose if you want to “move here”, “copy here”, “link here” or “cancel”. First of all no other menus in KDE have a “Cancel” item. Why is it needed here.
Second, the number of links you see on an average unix file systems is probably well below one percent of all files, and most of them is probably created by some script or install program. The only action that actually fits in the desktop metaphore is “move here”. This is also the most frequently used action.
So, in short every time the user drags&drops a file he will be asked chose from one very frequently used action, and some that he will almost never use. It would probably be better to do this similar to how its done in Windows Mac or Gnome, and some other desktop environments. That way new KDE users could take advantage of knowledge they might have gained from using other systems.
Another thing the art work need to be fixed. Not that it doesn’t look good in 3.4. It really does, but it sometimes clashes with the usability. Just look at the default desktop image. It contains a lot of KDE branding that makes it harder to identify icons placed on the desktop. At minimum the background image should be free from texts. Preferably the smallest image detail should be larger than the icon size. We can of course not prevent users from having whatever background images they want but the defaults should be selected more carefully.
Look at the default configuration of the pager in the kicker. It uses a 2×2 layout. It would be much better to use a 1×4 layaout that way the various desktops would be easier to hit (Fitts law).
But as you ask, here’s my list of what could be removed:
– Home: by default it just takes you to your home directory, the Home panel button already does that
To make it even worse the home icon is different in the konqueror toolbar and in the window.
Why is the default position of the trash not in the kicker instead of in the desktop where it often is covered by other windows. Both Gnome and MacOS have moved the trash away from the desktop and from eraly screenshots of Longhorn Microsoft is moving it away from the desktop as well. Why not follow that trend.
Actually, in KDE 3.4 you can do that, of course, the applet isn’t there by default.
Why is the default position of the trash not in the kicker instead of in the desktop where it often is covered by other windows. Both Gnome and MacOS have moved the trash away from the desktop and from eraly screenshots of Longhorn Microsoft is moving it away from the desktop as well. Why not follow that trend.
Actually, in KDE 3.4 you can do that, of course, the applet isn’t there by default.
And good defaults are everything. KDE3.4 could be a much better system without changeing a single line of C++ code, but just changeing the defaults a bit.
Well, Amadeo already gave you the answer I would have
You see, my problem is that I don’t think that those extra buttons stand in the way of productivity as long as the major functions (back, forward, reload, etc.) are clearly recognizable. For the average user that is. While reading your reaction I had the feeling that you reject everything that is not present in other browsers even if they make a lot of sense. Case in point: the clear location button. That button should be in every *nix browser, especially if you have to copy&paste links. It does not distract in any way really – does it make more difficult to write an address in the location field? However, with *nix like 3rd mouse button c&p it spares you at least one step (and it is a large one: you don’t have to touch the keyboard just to delete the previous address).
Well, I don’t want to go over each of your points (although I could, except for the security button – I agree on that point, but again: does it really hinder you?) – amadeo already did, but my point is that this ‘clutter’ issue was and is being blown out of proportions. Currently the nOOb’s major problem with KDE is the overwhelming number of apps it offers for various tasks. You need exactly two multimed apps: JuK and Kmplayer. But you have to install kdemultimedia (and noatun, arghhh) just to get a mixer on your panel – now that is a problem I think, but that problem might be alleviated by vendors and not necessarily the project (in my opinion – and the freebsd kde team is a good example: separate Juk w/o kdemultimedia, and ermixer for your panel:)))
>every time the user drags&drops a file he will be asked chose from one very frequently used action,
>and some that he will almost never use. It would probably be better to do this similar
>to how its done in Windows
This is pure usability, and a point where KDE are lightyears ahead of windows. In windows you never know what acction you get when you d&d, since it depends of what kinds of places source and target are. To know the result of the drag and drop up front you have to know the types, harddrive, networked storage or perhaps removable media. Different combinations yields different results, not what I call usability. KDE on the other hand are identical each time(You don’t get move when dragging from readonly tho)
The menu interrupts the users natural flow of work. This is an absolute no no in usability. Imagine sorting images from your digital camera into a couple of folders. You don’t want to be asked what to do every time. This is perhaps not much of a problem to advanced users that know how to handle key modifiers to expand selections, but it will certainly be irritating to newbies who will get this menu in their face much more frequently.
A better solution would be to always make a move when draging& dropping, regardless if the new location of the file is on the same disk partition or not. That way you would have the consistent behaviour you ask for, without unecessarily interupting the user.
Always move will not give you consistent behavior, it breaks the first time you drag something from a read only location.
Besides according to your example KDE’s way will force newbies to work more efficent by learning to move more than single files:-)
Always move will not give you consistent behavior, it breaks the first time you drag something from a read only location.
True, but displaying a menu would not make the origin less read only. Of course you could show a menu where the move item was inactive.
However, I really don’t think this is a problem if the origin icon had a read only ornament. People would realize that the read only origin was not going away. To make it even more clear you could get a warning dialog the first times you try to do something like that.
Besides according to your example KDE’s way will force newbies to work more efficent by learning to move more than single files:-)
Perhaps, but how do you expect them to learn that? Some will learn it from friends, other will go for the command line. Some motorically challanged people will have problems coordinate their movements enough to learn. The majority just get annoyed and switch to some other desktop environment, and this is a waist of otherwise very impressive technology.
The sad thing is, that even if you could somehow show that the menu had some minor advantage, I’m not quite sure it is a good idea just because of the simple fact that most people have been to exposed to the menu less way of doing things.
You can probably show that the DVORAK keyboard makes you type a little faster, but the difference is so small that it doesn’t compensate for the trouble of learning it. You could probably find a better user interface for a car than we have today, still nobody is changing it just because of the large userbase that would have to relearn.
The same thing goes for desktop environments. If KDE 3.4 had seen the light of day 15 years ago it probably would have had a chance of convising users that the copy/move/link/cancel menu was a good thing, but the window of opportunety for that is long gone. Things that doesn’t vagely resmble Mac or Windows (i.e. no drag&drop menu and double click for activation) will probably have a hard time to penetrate large user populations.
The general apperance of the computer desktop is allready defined in the minds of the users.
All those usability comments and analyses are starting to scare me. If you follow them, remove every button you can find (except if you happen to spend more than 90% of your computing time clicking on it), empty every menu, remove desktop images because they distract you from what you want to do…
Now what do you get? A perfectly non-cluttered, totally usable empty desktop. Staring at your blank screen, nothing distracts you, you can start to be productive…
Come on, people, if you you want to actually DO something with your computer, you need something on it. 🙂
If relaxation is really what you’re looking for, why the hell are you wasting your time on a in front of a computer? Get a plane, a car, a boat, anything, go to Kyoto and visit some zen gardens (which are, BTW, much more beautiful than anything you’ll ever see on your screen).
On a more serious tone, I do agree with Ugo Engborg ; the KDE team need to focus on offering better defaults. But the 3.4 release did show significant improvements, I hope to see much more in the 4 release.
Until then, don’t forget that you can configure toolbars (and virtually eveything else as well) in KDE, making them as uncluttered as you want to be. And is is NOT hard to do. I can do it, believe me, it’s not hard.
I’m getting tired of people criticizing KDE when they haven’t really tried to use it, or when they only have opinions which refers to the 3.1 release… Things have improved dramatically in the last releases (speed, better UI). Spend less time analyzing screenshots and more time actually testing KDE, I do believe some of you might be surprised.
Only geeks configure toolbars. Only geeks tweak and play with configuration settings. That’s why it is important to provide useful and reasonable defaults.
A web browser in my opinion should not have more than seven elements in its toolbar. The go, back, forward, stop and refresh buttons, the location bar and the browser spinner.
Those are the most useful and used elements on the toolbar. Every other function can be accessed via the menus. Advanced users, if they wish, can go ahead and add as much crap as they want to the toolbar.
It really makes me smile when people complain about d’n’d fo newbies…newbies DON’T EVEN KNOW what d’n’d is. d’n’d is a HIDDEN feature, the natural workflow on a mouse button is “click once and releas after a while without moving mouse”. Drag’n’drop is used by intermediate and maybe advanced user, while power users usually use keyboard shortcuts.
Or maybe we give a different meaning to “newbie”
Quote:
“Only geeks configure toolbars. Only geeks tweak and play with configuration settings. That’s why it is important to provide useful and reasonable defaults. ”
Not exactly true.
A normal user configures toolsbars and plays with configuration settings too.
Your “solution” is not a solution at all, you only move the problem.
Better would be to make configuring toolbars and settings as straightforward as possible so that it takes no time at all to even understand what you’re changing.
Defaults are just that, defaults.
I have NEVER, read: realy realy NEVER EVER installed for example a frequency controller for a motor and used its default values. NEVER
I have NEVER, read: realy realy NEVER EVER bought a default car, I always got extra options here and there, or left out options that I didn’t need.
It is not the task of KDE to come up with a default that works for everybody. That is just not possible.
What you think are good defaults, another person will think they suck. What should KDE do in that case? Include both? leave out both? include one and not the other?
Or should it provide an easy to understand framework so users can adapt their KDE to their needs?
tb,
Can you point me to any member of your household who isn’t a geek who actually reconfigured Internet Explorer’s toolbar? Once again, only geeks have time for that crap.
My suggestion does not hide any problem. It merely advocates that the most used and useful functions be presented to the user as defaults.
Functionalities that are hardly used should be removed from the default interface. The is not a question of pleasing users, this is a matter of rationality, sanity and usability.
The problem with KDE today is that it innundates users with too much information, functionality and options at the interface level, 90% of which are useless. Information overload leads to a variety human ailments, such as stress, anxiety, confusion and even nervous breakdowns.
That is why Pilots in training are not immediately thrown in front of a 747 cockpit. They’ll freak out. They begin with a cockpit that has a spartan interface and then gradually move up to more complex cockpits.
The same applies to GUIs. Begin with a spartan GUI with the most used functions, then as users become advanced, they can manipulate the GUI to suit their needs. KDE does it the other way round. Through as much crap at the user as possible, they’ll figure shit out themselves.
Every time I have used KDE, I have had to waddle in configuration hell, searching for that options, tweaking that function, adjust that look, etc. It doesn’t help that the tool for adjust toolbars in KDE is pathetic.
In an ideal environment, one wouldn’t need to tweak anything except to unvail advanced or peculiar behaviours. It is just absurd to find a menu with 20 items only to observe if one does a survey, 90% of the users only use 4 to 5 of the items in the menu.
many people complain when people are suggesting to have these tings removed by default. they say i use x,y,z featre and its so easy to remove so why remove it by default. wel if it’s so easy to move it should be really simple to add. and persons that want more feuters are often better at tweking.
If you want to go to http://www.avl.com by only using the mouse (if the text is not hyperlinked which happens quite often in comments like this), you have to copy the url into the URL bar.
The most efficient way to do that in mozilla would be:
1) select the current text in the URL bar and delete it.
2) select the new URL.
3) press middle mouse button in URL bar to insert.
4) Hit return (now you needed the keyboard anyway)
This is totally counterintuitive, because if you want to go to http://www.avl.com you would first select the URL, then move the mouse up to the URL bar, and now you need some way to efficiently delete the current text in the URL bar. Konqueror provides one, mozilla does not, therefore this button is not ueseless, I need it quite often, and I think the way I use it is the real reason why its there.
Geri
Actually with the various textlinks extensions for mozilla and firefox (and Opera does it natively too), you can just select the text link and click on “go to url”
It really makes me smile when people complain about d’n’d fo newbies…newbies DON’T EVEN KNOW what d’n’d is. d’n’d is a HIDDEN feature, the natural workflow on a mouse button is “click once and releas after a while without moving mouse”. Drag’n’drop is used by intermediate and maybe advanced user, while power users usually use keyboard shortcuts.
That’s true. This is yet another reason why the drag&drop menu is not a good idea, especially as it contains things that can’t be done in any other way.
The only way to create a link is by drag & drop.
Not that this is such a big problem to a newbie, as he will probably not know what a link is, and linking is an extremely rare operation even for experienced users.
On the other hand, if the newbie doesn’t know what a link is, and the experienced user makes links very seldom it shouldn’t get in their face every time they does more common operations like moving files.
Also notice the advanced user have no way to know that he is making a soft link (as opposed to a hard link), and the GUI offers no way of creating hard links.
Or maybe we give a different meaning to “newbie”
I think we do. Today it is almost impossible to get test subjects for usability research that haven’t been exposed to some kind of desktop environment. Here in Sweden you typically get people younger than 7, or older than 60.
Unfortunately this doesn’t mean that all of them know about modifier keys and other shortcuts. But most of them would probably have seen drag & drop in action.
To handle real newbies in the sense you seam to think about them, you probably need some kind of interactive tutorial to run automatically the first time you run the system.
Early version of MacOS contained such tutorials where the user learned what the mouse is, how to drag&drop, click, double clickt etc. Adding a tutorial like that would probably be a good idea for KDE, Gnome and other free desktops.
I don’t know why everybody is so focused on newbies these days. The reason I bother using linux and KDE is that it provides me with maximum flexibility and control. If I want something easy, I would probably choose a mac. If KDE becomes a newbie oriented desktop like gnome, I’d be the first to switch to something else.
As far as a default setup is concerned, I think its importance is overrated. A user would see/use default exactly once. Once you configure kde the way you are most comfortable with, the setting will be there as long as you keep your home folder. I have not seen a default kde login on my computer ever since the kde 3.1 days.
Well said. If you follow the debate b/w Uno Engborg and Morty one thing should be clear: there seems to be a market for both the GNOME and the KDE approach – for that’s what this is all about. What I find really funny is that my experience often tells exactly the opposite these “usability experts.” Also, if you examine their rhetoric, you’ll find strong declarative sentences without clear elaboration (why should the clear location be removed? It’s confusing. Well show me a person who is confused by it
))) – and they will, you can bet on that!
I have a computer-agnostic girlfriend (he thought her laptop is running OfficeXP) – and she was never distracted by buttons in KDE. She quickly recognized those that she needs, and developed a blind spot for all the others. Of course, a UI can prevent that if it is very very cluttered – but this hasn’t been true of KDE since 3.3.x times. There is a single icon I don’t find useful: the security. And I haven’t even noticed it or gave it a thought until someone here draw my attention to it. So I will repeat it: this clutter thing is blown way out of proportion. Not only that, but in the vein of GNOME’s “less is more” philosophy, some folks would have KDE remove all those useful features that its users like. Not only that, but they will blindly denounce some things – in the name of usability no less – that should have been implemented on *nix like d&d systems long time ago (clear button!).
We can engage in a war about numbers – more ppl. use KDE than GNOME and such things – but lets just say, and I think everyone agrees with this, KDE and GNOME has more or less equal userbase. And the usual flamewars that follow every comparison is a good indication that we need both! I will hate KDE if its developers would listen to _some_ usability pundits (note the emphasis, there are exceptions of course) instead of their users (or their own usability team!) and remove everything that makes it a lot more usable than competing desktops. And I could go on citing a long list of examples where GNOME for instance consciously violates its on HIG – but I won’t b/c their usability is not my usability, and my usability is not a typical ‘geek’s usability either. Oh yes, how I hate that: whenever I argue for the usefulness of a feature, I get, yes you are a geek, not the average user – yes, and neither is my girlfriend, or students visiting our library who find no problem in using KDE to do their works
The point is: there is no Holy Grail of usability. There are various usability studies/approaches, and KDE’s is a perfectly valid one, just like GNOME’s. Neither is closer or farther from the proverbial Average User’s expectations – they are different, that’s all.
I don’t know why everybody is so focused on newbies these days. The reason I bother using linux and KDE is that it provides me with maximum flexibility and control.
The discussion isn’t about removing functionality, flexibility or control. It’s just about defaults.
If I want something easy, I would probably choose a mac. If KDE becomes a newbie oriented desktop like gnome, I’d be the first to switch to something else.
So you want KDE to be complicated so that you can feel superior over newbies/lusers/whatever?
Simplifying the defaults wouldn’t stop you from optimising/complicating it any way you like.
As far as a default setup is concerned, I think its importance is overrated. A user would see/use default exactly once.
If default setup was so unimportant, why are you complaining about demands to simplify it?
In fact, of course default setup is very important, because it’s the first impression any user gets of a program and because it’s the only setup many users will ever see.
That’s because they simply aren’t interested in tinkering with program settings or are too scared to break something. And even experienced computer users initially won’t know what can be changed and how to change it.
I don’t know why everybody is so focused on newbies these days. The reason I bother using linux and KDE is that it provides me with maximum flexibility and control. If I want something easy, I would probably choose a mac. If KDE becomes a newbie oriented desktop like gnome, I’d be the first to switch to something else.
The problem is that if only a few experts use KDE, no market will be built around it. In the end that will mean that your boss will make you run MacOS, or some other in your oppinion dumbed down OS at work, as he will not see any business value in training his employees to get your level of skills.
As you point out, configuring the system is something you do very seldom, and this is probably not much of a problem to advanced users. If the default settings make KDE difficult for new users, they won’t use KDE. To them the defaults are important.
The time of the macho “if it can’t be done in vi its not worht doing” attitude is long gone.
Twenty years ago a bank clerk could tell a customer that a service a customer needed wasn’t possible because of the computer couldn’t do it, and that would be the end of it.
Today, the customer asks why.
In the bad old days people humbly accepted the word of a computer guru, and the guru held his position as guru by using difficult tools and the people around him admired him for mastering these tools. Nowdays, people just say: hey, its a computer, what kind of a developer are you if you can’t make it simple enough for me to use.
Things really have changed havn’t they.
Well said. If you follow the debate b/w Uno Engborg and Morty one thing should be clear: there seems to be a market for both the GNOME and the KDE approach.
What I really would like to see is a third approach. Something that looks, and functions as simple as Gnome in its default configuration but is highly tweekable like KDE.
A new user should not need to know that he runs KDE or Gnome or something else. The desktop methaphore is now so old and settled that all desktops should share some minimum feature set that works more or less alike regardless of the make of the desktop.
A person who can drive a Wolkswagen imedialtely recognize the controls in a Ferrari and is able to drive that as well, or any other car, with the possible exception of cars made in France. Why can’t computer desktop environments work like that.
Today KDE is a little like a French car. Their followers love it and the rest hates it. This polarization isn’t a good thing if we ever is going to see a wide spread use of a free desktop of some kind. To meet future products of Microsoft and Apple cooperation is needed. The freedesktop.org is making wonders here, but more is still needed.
What I find really funny is that my experience often tells exactly the opposite these “usability experts.”
If that’s supposed to refer to my posting about removing konqueror buttons: I never claimed to be a “usability expert”; as clearly stated, it was a personal opinion.
That’s of course the problem with usability discussions: it’s largely a matter of taste and opinions, because scientific studies into that area are difficult and expensive to do, and even then the results aren’t necessarily conclusive. That’s no reason to stop arguing about it though
Also, if you examine their rhetoric, you’ll find strong declarative sentences without clear elaboration
But your sample of one girlfriend is meant to be any more significant?
(why should the clear location be removed? It’s confusing.)
I didn’t say it’s confusing, I said “Who ever uses it?”, implying that it’s not useful enough to warrant inclusion in the default toolbar setting. Right-click->Clear already does the job, and works consistently for any textbox in KDE.
Granted, one extra button doesn’t make konqueror confusing, but a lot of them together certainly do. Well, at least the cut/copy/paste buttons have gone in 3.4.
I don’t like the GNOME approach of actually removing functionality or hiding it in something like gconf either.
I think KDE should be both powerful (which it is) and easily configurable (which it is to a degree, at least compared to something like The Registry), but the default settings should follow a minimalist approach in order to make it easier and less intimidating for newbies and to provide a clean slate for geeks to base their own configuration on.
Besides, an uncluttered interface simply looks better, but that of course is entirely a question of taste.
I very much agree with you, but the cars are called “Volkswagen”
I’m sorry if I came of ranting
Actually this is one of the best debates I’ve seen for a long time on osnews – and that goes for everyone I don’t necessarily agree with: you, Engborg, etc. My problem is that I see the same reasoning repeated again and again without taking into consideration the changes that took place in KDE. I think KDE presently it’s ‘there’. Where ‘there’ is? I don’t really know – it’s not ‘there’ at the level of the ‘Average User’ – which imho does not exists. I’ve seen non-geek users (my 60yo history teacher, the librarian here, etc.) who actually spent considerable time to ruin the usability of their desktop 
– or at least it was a ruin for my concept of usability: putting everything but the kitchen sync into microsoft’s office toolbar for instance, even though they had icons on the desktop as well. What I mean by ‘there’ is simply that I don’t think there is any outstanding bug in – at least not in 3.4 – that stands in the way of productivity in any noticable way. In fact, some features that other DEs currently lack actually can help new and power users alike.
That doesn’t mean that I think of KDE in black and white terms. Currently its biggest problem is duplication/arrangement of its tools. This is not an easy issue, for one can argue that it is the responsibility of various KDE vendors to customize the default layout. Problem is they don’t do it
The old example is the 3 text editors. I understand perfectly that they serve different functions – but still, I think kate should be moved to sdk or kdewebdev. Or kdemultimedia is another bad example.
Granted, one extra button doesn’t make konqueror confusing, but a lot of them together certainly do. Well, at least the cut/copy/paste buttons have gone in 3.4.
I absolutely agree. Users are already familiar with copy and space shortcuts.
Anyway, I’m sorry for lashing out unnecessarily at you or others
Quote:
“What I really would like to see is a third approach. Something that looks, and functions as simple as Gnome in its default configuration but is highly tweekable like KDE. ”
Yeah, that’s what most people want.
A clean desktop, but thousands of possibilities to tweak.
They have to be tweaked somewhere.
And that’s the big problem today… how do you make it easy and straightforward to tweak thousands of settings.
Some suggest good defaults, but that’s only a small patch to a big wound. Defaults only work for a very small target audience. Some people like to play a lot of media files, and like to have control over them via some toolbar buttons. Others don’t, and do not want to have those buttons placed on toolbars in various programs.
In the end, you’ll be creating tens of defaults for each target audience; new users, moderatly new users, advanced users, media addicts, developers, …
Some things in KDE can be done better though.
And the developers are aware of that, just check out http://dot.kde.org/ Read the blogs and the news postings there and you’ll see that the problems are known, but that there are good reasons why some of them are not adressed yet.
And to reply to a question asked earlier, if I know someone in my family who is not a computer geek and who has changed the IE toolbars:
Yes, my brother in law. He doesn’t know anything about computers, but loves to surf on the internet, and he did change the IE toolbar.
I tried to convince him to use Opera or Firefox, but he won’t because he doesn’t know how to use them.
Believe it or not.
Re:
Yeah, that’s what most people want.
A clean desktop, but thousands of possibilities to tweak.
They have to be tweaked somewhere.
And that’s the big problem today… how do you make it easy and straightforward to tweak thousands of settings.
This is not as much of a problem as you might think. First of all KDE allready make a very good job of making all these settings easy to find considering how many settings that actually can be made. E.g. the config dialog is searchable.
Second, with good defaults the number of things a user have to config could be limited. To make it even esier there is help available even though the wording of the text could benefit from some improvements. Making things configurable is something KDE allready is good at. What it is not good at is selecting the defaults.
Look at the kicker panel where, you don’t make use of the screen borders as much as you could by making it unnecessarily high, and have multirow panel applets for the pager and the activity field. Gnome do use two panels for a reason…
And why use single click for activation, when most of the potential users will be used to using double click. You can argue all day that it will reduce carpal tunnel syndrome, and that it might be easier to use by elderly or motorically challanged people. However most users don’t like it perhaps because it will make file selection harder, or simply for the reason that they are not used to it. There are in fact a lot of studies that indicates that single click should be better but they usually involve less complex situations than a general desktop use, e.g. information kiosk and such.
However, Microsoft tried to introduce single click in their active desktop a couple of years ago, but it never got widely used. KDE is not likely to have any more success. You could hardly get a larger field test than the Microsoft Active Desktop, and a more similar use situation. Even though Microsoft failed, KDE still insists on single click as default.
Why is the KDE control center placed in the appliation section of the K-menu. The things you configure here is orthogonal to all other applications. The kcontrol would have fitted much better in the action section or as a button on the kicker panel. Neither does the home directory menu item found in the application menu belong here.
Another thing, that have been discussed on the KDE usability list is to use one MacOS like application menu as default. From a usability and KDE only perspective this is a very good suggestion. It saves screen space and menus along the screen border is very easy to reach as you can’t over shoot.
The problem is that, to make this a good idea, they need to bring other desktop environments on board so that e.g. Gnome and Wx-widget applications behave the same way. Mixing multibple menu models from various toolkits will not be a good idea. If KDE developers could make this the standard e.g. through freedesktop.org it would be a good thing.
GNOME has money from Novell, yet they conduct no usability studies. It’s too bad, because C# (Mono) has a brighter future for development than C++ (KDE), in terms of development speed (undisputable fact inferred from C++/Java comparison).
However, GNOME doesn’t care about usability. You can’t even edit the menu with hacking some obscure file. KDE, OTOH, has been consistent in providing an easy interface experience. GNOME is locked in their Mac OS 7 interface guidelines. Their hacker ethics and C macho habits will prevent GNOME from delivering. The situation is probably near helpless, since they do not even look at the patches people sent them in 20% of the time (this statistic appeared here in the Eugenia vs GNOME controversy).
I guess it would be far to say that part of the success of KDE has to do with a language that was designed for better abstraction and modularity. Phillip Greenspun says something that appears to be apply to GNOME: if you C project is too big, C proves to be not so fast; and we all know how GNOME is slow in launching apps.
It would be nice to have some KDE people consider C#. I don’t know how this is going to happen, due to their dependance on Trolltech’s Qt. But it would seem like a smart move toward even faster development. Bindings would be nice.
I’m glad to see that there are people in KDE that worry and specialize in user experience. I guess that would account for the fact that, whenever you go to a huge F/OSS event, you see a lot more of KDEs UIs than GNOME. I mean, the wife doesn’t complain with KDE, and that’s *my* usability experiment 🙂
The speed of development on KDE does not only come from C++, but to a big extent from the features Qt adds to the language. Bindings for C# would be nice as you say, but I doubt they would have any big impact. Languages like C# and Java don’t offer you that much advance against the C++/Qt combination, neither in development speed or maintainability. As a example you can take the Java bindings for KDE, they have been kept up to date for every release since about KDE 2.1. But they do not seem to see much use.
If you want faster development for KDE, Ruby or Python are much better choices. They are much more suitable for RAD development than C# or Java, and the KDE/Qt bindings are mature and stable. A recent comparison between C++ and Python development with Qt, showed a average reduction in lines of code with over 40%. http://lists.kde.org/?l=pykde&m=111279032226911&w=2
> Stuff like search-based filing doesn’t really interest me,
> but this link engine is an idea I’ve never heard
> anyone mention before.
Microsoft plans to include this in Longhorn (or a later version) too.