OSNews.com is featuring an interesting three-way interview (5 pages) with KDE's very own Waldo Bastian and Aaron J. Seigo as well as GNOME's Havoc Pennington. An interesting diversity of opinions on various UI and usability issues is presented. "On the Ok/Cancel issue: KDE has implemented things in such a way that allows all KDE applications to have the order of these buttons flipped with a change to a single line of code. It is exactly this sort of brilliant design that allows KDE to be so internally consistent. So the question often comes down to whether or not we should make a change, rather than if we can. Personally I think it is irresponsible to impose personal aesthetics on your users in a seemingly random fashion by disrupting the interface they know without very compelling reasons to do so."
Comments
> I think that KDE should spend time on porting just the kdelibs on the other qt platforms,
Sure, you start.
> Sure, you start.
Someone should sponsor that - one who has financial interests, and in our case it's QT. Then the benefit is for both QT and KDE.
There is an OS X port already. Of course, the usability is /very/ limited since most of the GPLd apps have an unclear legal status wrt to linking to the commercial edition of Qt; but this does mean that an app using kdelibs can now run on 2 platforms.
"But if they would precisely calculate the total costs, they would see that they NEVER get their money back. The costs of Qt developer licences are VERY small compared to the gain of productivity."
This is the exact same argument that Microsoft uses to claim that they deliver better value. Don't get me wrong; I agree with everything you say, it's just ironic. (though it's *possible* of course that Troll could become the new MS some years down the road...)
Not really... It's GPL'd. We have the source, and anytime they decide to take things in a direction we don't like, we can fork the thing.
At the same time, in the past and at the current moment, they have given us no cause for this. QT is good and it keeps on getting better.
The GPL is a great safety mechanism for exactly this sort of thing :)
Is it GPL on Windows & Mac and PDA's?
One of the reasons to use QT is that it's cross platform. If only one platform has an up-to-date GPL'ed version it severely hinders the spread of QT GPL apps. Give me Konqueror for windows!
QT is GPL on X11, so you can use it on any platform that runs Xfree86 or another X11 implementation.
There is a port of KDE/QT on Cygwin :
http://kde-cygwin.sourceforge.net/
"KDE on Cygwin is the port of Qt and KDE to Windows using Cygwin, the POSIX emulation layer for Windows, and the Cygwin XFree86 server."
You can also patch QT/X11 to run without X11. There is attempt of a native port of QT2 on Win32 :
http://kde-cygwin.sourceforge.net/qt2-win32/index.php
Thanks for the info. But I'm afraid this does not address any of my concerns:
Uptodate != There is attempt of a native port of QT2 on Win32
Native != (Win 32/Mac OSX) + X11
I dont use QT in my GPL projects because of this even if I would love to.
I think its stupid of TT not to have GPL versions on all platforms. Think of the free bug testing/fixing they would get from people porting GPL QT apps to other platforms.
Yes it's GPL:ed so there will be no problem for free software. The issue is commercial software. If KDE becomes a de-facto standard, like windows is today, Troll could in theory charge as much as they pleased from anyone developing desktop software.
This is one reason that keeping GNOME alive is probably a good idea (after all).
As I had pointed out on slashdot a day ago, I tried Gentoo linux just to get the best KDE 3.1 install I would like. One of the options in Konqueror confused me hell. The first setting in the settings menu on konqueror is 'Show Menu'. It is awfully easy to click it. Then we get confusion. Where's the menu?
Sure someone pointed out that the option appears when we right-click on a web page in konqueror. But I did panic a little and tried every key combination I could. Finally I got ctrl+M.
My father also uses the system (which dual-boots with WinXP incidently) and gets highly confused and panics in WinXP. He likes KDE and KMail esp. I would like to see what would happen if he hits this option (he might bring down the house though).
Seriously guys, take this option off the menu and place it in the toolbar menu (rt-click on toolbar). At least it remains there when the menu disappears.
Please keep options that make their access point disappear away from that access point. Don't make a toggle option disappear from one place or access point and appear in another. It confuses users and frustrates them, especially if the option is accessible easily by mistake. I'm on the look out for other such options and will certainly report them. Right now I'm reinstalling Gentoo since my disk space estimate was way under my requirements.
Um, you're right. "Show Menu" hiding the menu is bad. It's not at all obvious that it will hide the menu when you click it. FWIW, you get the Show Menu option in your context menu once you hide the menubar.
"Please keep options that make their access point disappear away from that access point."
Very good suggestion.
Agreed. I've seen way too many people get confused by this on IRC, as well. I think at the very least adding the entry to the toolbar edit menu would make this easier to get back. Not sure about removing it from the menu[1] entirely (it might even violate the style guide to do so, IIRC), but this should definitely get thought hard about. I'll try to make a patch for this, I guess, if I don't forget. I think I should build kdepim or something so I don't loose my TODO entries....
[1] Although arguably, hiding memunar is used the most with the fullscreen mode, where you can't get to the menu anyway since it's hidden.
I also think it should be "Hide Menubar" instead of "Show Menubar" (or toggle between the two depending on whether the menubar is being shown or not).
The three major features of any GUI that provides access to a program or an OS is depth, breadth, and focus. Breadth defines the gamut of features that can be handled by the GUI. Ideally every useful function of the serviced software should be handled by the GUI to have a breadth score of 100. Depth defines to what level the individual functions can be handled by the GUI before you have to resort to the command line. KDE is solid and moving forward rapidly in both of these areas.
Focus is the hardest of these functions to implement, because its form can vary depending on the user. A good example of a program UI with great depth, breadth, and almost no focus is Emacs. If you want to browse the web, edit binary files, or write a TeX document Emacs is probably a good choice, but who would recommend it to a newbie who wants to edit a short note?
GNOME has better focus than KDE but for the wrong reasons. It simply offers less choices. It is easy on the eyes and simple to find the basics, but once you get past those levels the options are limited. The trick is to configure the initial interface to be familiar and understandable to the newbie, and easy for the advanced user to easily find and use the more complex functions. This is difficult.
The sort of menu selections you describe are a focus issue. Before placing any item on a main menu remember the purpose of the program, and decide whether this function enhances or degrades its functions and usability. KDE would do well to concentrate on what should be on initial menus and what should be in advanced config dialogs. The desktop has a wealth of powerful and reliable features, but they are often arranged in a confusing manner. I agree that the Red Hat 8 rationalization is a mistake. GNOME is the slower ship in this convoy, and you are forcing everything else to its speed. Let them sail alone.
Windows XP is a good example of MS redesigning the user interface to be easier for newbie access but very good at offering all of the features. Win98 and ME were UI disasters where focus is concerned, with program menus scrolling off the screen and toolbars bumping toolbars on the task bar. They obviously got some feedback from the focus groups.
I am finally moving some of my desktop work to Linux, after 4 years of using Linux as a server and Win as the client platform. OOo, KDE 3.0+, and the work of many open source programmers has brought Linux to the point that it is a viable desktop option for experienced users. It still needs work to get to the usability level of Windows for the new or occasional user.
I am very impressed with Qt and the modularity and mobility it seems to offer. I am very happy to see that the Kylix developer package (I use Delphi in Windows) provides access to Qt though a wrapper level. I am looking forward to seeing how it works.
-ronnie-
I've done that before myself. I was not impressed.
Rather than removing the feature, it could pop up a dialog box at the same time.
"You have just removed the programs menu? Is this what you wanted to do? Hit Cancel to bring the menu back. You may also press ctrl-M to bring the menu back later"
Also have a "click to dont show this again box"
Corey
just to add to this, though offtopic.
I found backing up of kmail (with maildir) very tough. Since I didn't have any critical mails, I deleted it in order to repartition my drive (I have a new 80Gb barracuda disk to play with:-). There should be an option to export kmail mailbox to mbox format so that it is easy to backup.There are only import functions currently available. Or have I missed some other 'obvious' method to backup ~/Mail to a fat32 or fat partition?
Earlier versions of kmail ~/Mail folders were quite easy to backup, just copy them to the other safe partition (usually fat32) then copy back again. Now it doesn't work.
How to backup ~/Mail (on Linux):
cd ~
tar cvfj Mail-backup-2003-03-12.tar.bz2 Mail
Then copy Mail-backup-2003-03-12.tar.bz2 to your backup medium (say a fat partition).
The above is the standard way to backup Unix directories to fat partitions since this is the only way to make sure that you won't lose any file attributes (owner, permissions, etc.).
How to restore ~/Mail (on Linux):
cd ~
tar xvfj Mail-backup-2003-03-12.tar.bz2
How to export a folder to an mbox file:
1. Select the folder
2. Press k to select all messages in this folder
3. Press Ctrl+s to save the selected messages.
The selected messages will be saved in an mbox file.
"Most laughably, I could not move a music file from the desktop while it was being viewed in the file manager."
That's not a bug, that's a feature of Windows (and not a new one either).
Feature? Call it what you want but this is definitively a bug.
Windows can lock files while they are accessed or in use much
like Linux can. But more often than not Windows Explorer
locks up files only when they are selected most probably
because a preview is shown and Microsoft "forgot" to
release the file. It's obvious this
is a bug and IMO a pretty major one which hasn't been
resolved for ages.
Wrong,
File Locking is a feature.
Locking a file to generate a preview, displaying the preview, but then not releasing the lock until it gets cleaned up when the process dies (or a different image is selected), is a bug.
I've been caught by that one too, and it's really irritating when it happens, not to mention completely unintuitive if you don't understand the concepts of file locking.
"On the Ok/Cancel issue: KDE has implemented things in such a way that allows all KDE applications to have the order of these buttons flipped with a change to a single line of code. It is exactly this sort of brilliant design that allows KDE to be so internally consistent. So the question often comes down to whether or not we should make a change, rather than if we can. Personally I think it is irresponsible to impose personal aesthetics on your users in a seemingly random fashion by disrupting the interface they know without very compelling reasons to do so."
Actually exactly regarding on this topic there is a little issue, with Kate (and may be with other kde programs - thanks to that consistency), it's about the dialog that appears for unsaved file when closing kate. It seems that in kde 3.1 the 'No' button is replaced with 'Discard' button, which is very confusing for the first time it happens, but the real problem is that for people like me that are using shortcut keys for faster dealing with apps - the shortcut key is changed too (before for removing this dialog without save i used Alt+N (&No) and now it's Alt+D (&Discard), ESC is not doing the same job because it just makes the dialog to disapear).
So the point is for kde at least to be consistent with the previois versions at least for shortcut keys, because old fans of kde that have learned to use that keys instinctively, will have their work slowed and at least little disapointed.
Tip: add configuration option somewhere or just make the old combination work.
PS: I appologise if this conf option exists somewhere, so if someone knows where it is - just tell.
Congrats KDE you are doing a great work.
Sorry. I do not agree. When I changed from Windows
to Linux some Key-Combinations (F5 in Mozilla)
just did not work. I was annoyed at first but in the mean
time I got used to it. "Discard" is by far more appropriate than "No"
and I guess that's why they changed it. They eye usually
only scans a dialog, just looking on the buttons and not on the question.
If the buttons are labeled more descriptive you can click directly
without reading everything. And it's by far easier to mistake Yes for No:
Do you want to save the files before you quit?
Do you want to discard any changes?
I've actually seen both questions in various applications.
Look at "Run Command". The button is labeled "Run" and not
"Ok" for the same reasons. Usually I'm in favor of configuration
options but this would be really bad: Options for every single
short-cut key - I doubt that's the right answer. In this case
you better get used to it. There is always a trade-off between keeping
things in place for the user base and improving things.
My point is not with the name of button, I dont care 'Discard' is understandable of course. It's not a problem to just click it, moreover it's on the same position.
My problem is what clicking it is slower than pressing Alt+N (which was before when the button was called New), now it's Alt+D. And for people who get used the old key shortcut it's disappointing that old shortcut does not work because it slows their work.
(I just need the old shortcut nothing more)
I think the more important thing is to make it obvious what choice you are about to ask the user to make. Something like:
You have clicked to close this document, Please select to:-
[button]Don't save[button] : [button]Save[button]
Most importantly, color code the buttons. Red seems to tell of impending doom. If you Don't save you will lose all the work you have just done. Green for go, save your work. I am assuming when most people work on whatever they work on, they want to save their work.
I have heard people complain about button order in GNOME and how it is confusing for the user, but I only realised it after reading it somewhere. And no, I have not used a Mac. The dialogs were good. The onus is on the guy who writes the program to make that intuitive and to enable people to understand and make the choice easily. I think interface guidelines for both KDE and GNOME ask developers to do this.
I also tend to think developers have to make choices for users of their products. I do not always buy into this 'choices' stuff. If you are a usability 'expert', I expect you to make choices for me the user that will make me more productive. These choices are made a lot in other spheres of life. Computers should not be an exception. Choices can exist. But they do not all have to be in one desktop environment. Look at Apple, they are confident enough in their default interface to lock it down somewhat and make it difficult to change and they still have every other desktop environment trying to skin theirs to liik like OSX. That is good engineering.
If you ar a devloper who cannot make these choices, then go and be employed by some interface designer who will use you just to code, and make money off you while you are at it.
Before I say anything else, it should be mentioned that the entire Yes/No, OK/Cancel thing was a very informed decision taken with direct user testing evidence, before 'Discard' was added. People do get very confused with what they are exactly agreeing to in such dialogs, and 'Discard' seems to alleviate that somewhat.
"Look at Apple, they are confident enough in their default interface to lock it down somewhat and make it difficult to change and they still have every other desktop environment trying to skin theirs to liik like OSX. That is good engineering."
Not really... You just hired some designers and made it look pretty. That's why people skin it. It is worrying, however, that people seem to have got it into their heads that all UI changes Apple produces are naturally perfect. Apples' products are pretty, not necessarily terribly usable. The entire idea of ejecting a CD by dragging it to the trash, for example, is utterly stupid, is counter-intuitive and breaks the metaphor in so many ways, it isn't even funny.
In general UI terms, I do find MacOS and OSX very dubious. The only point I find genuinely nice are springloaded folders (their general file browsing is horrible), but a slightly faster way to copy files does not generally benefit anybody's workflow (unless you copy files for a living!).
Purely in technical terms, Quartz is poor at rendering large numbers of windows... The memory footprint with lots of windows is horrific. Therefore, Macs only really are usable for people who run small numbers of apps at the same time, with limited window numbers. Similarly, the file management is good if you need to move one file, but bad for large numbers. These are serious usability issues, and poor scalability is indicative of poor engineering, in fact.
Sorry for the rant, but it just gets on my nerves when people talk as if Apple have the best UI design, and that we should follow it. Linux should pick and choose the best ideas, by all means, but it shouldn't copy verbatim.
I'm reasonably active on the kde-usability mailing list, and I think that the usability changes and cleanups that are being implemented at the moment will make your Linux box easier and more enjoyable to use in the near future, with new ideas and without copying the cruft :)
I do not say Linux should copy them, but our developers should not leave interface design to the user. Make choices and stick by them. Nothing enhances productivity more than a slowly changing, or even unchanging interface. A user could make a horrible choice and not know how to revert. That is admittedly a worst case scenario.
I have read many reviews by professionals, and they all sem to agree that the mac interface is the best from a user standpoint. But we are all entitled to our opinions. KDE does however seem to have more in common with Windows XP than with OSX. I am not saying they are copying. They chose a route and are sticking to it.
Look here for the so praised consistency of GNOME. The so high hyped 'clean look' of GNOME nothing more than a big joke. GNOME finally realized that KDE has won and now they are trying to fuck KDE up with the same stupid arguments they used to piss not a less amount of GNOME people off. As if it wasn't enough to fuck GNOME up now they are trying to do the same with KDE. Cooperating in a certain way is ok but this is getting to far for my taste. Havoc Pennington is responsible for the sad horrible failure that GNOME has nowadays.
http://www.osnews.com/story.php?news_id=2706
I wish KDE had a high profile person like Havoc. He is professional enough to raise above the stupid religious flame wars and see the big picture and listen to arguments and compromise.
http://lists.kde.org/?l=kde-core-devel&m=104725441423600&w=2
And I must say that even if Gnome lacks features I like what theyve done with Gnome 2, good GUI organization is better than clutter.
BTW.
Did you know that Trolls like you have a lot in common with Gnomes?
Fuck off offending dipshit. You are quite unfair calling people with own opinion 'Troll'. Investigate into the whole GNOME community and interview people and most of them belive that Havoc Pennington is a narrowminded and stupid person who only cares for his own vision of a Desktop. He doesn't care for other peoples opinion at all and he is responsible for the split GNOME community.
>Fuck off offending dipshit.
Oh, that hurt..
>Investigate into the whole GNOME community and interview people and most of them belive that Havoc Pennington is a narrowminded and stupid person who only cares for his own vision of a Desktop. He doesn't care for other peoples opinion at all and he is responsible for the split GNOME community.
Is Havoc responsible for this major breakthrough!
Did he really managed to get KDE and Gnome to agree on something?
If he managed to piss of separatists in both camps he must be doing something right.
See, the biggest problem that GNOME has is their time is running away, the powerusers are running away (those that could be your developers are developing KDE apps nowadays), your own people are not necessarily happy, Miguel de Icaza and Havoc Pennington have both admit that KDE is far better. Comparing the GNOME newspages you read all sorts of things like 'KDE and GNOME finally agreed on unifying the HIG', 'KDE and GNOME finally agreed to work together' and 'KDE and GNOME finally agreed on unity'. And beliving in this the same people announce this happy cooperation on various news pages only to suggest the readers that there are big things going on. While on the other hand reading the KDE pages shows the exact opposite. Even the unified HIG page is stagnating. Knowing the fact that GNOME has all these issues there are some poor attempts from Seth Nickel and Havoc Pennington getting the KDE developers into cooperation. But you are missing some points here. The KDE people are in no need to create standards because they already made a working consistent Desktop while GNOME still suffers from simple things such as Filechoosers. This kind of cooperation will only throw the serious development process of KDE significant back and re-implementing all the things the way GNOME developers like to see it will put the Desktop back for another 2-3 years on Linux. See, KDE are the first one with a working Desktop and KDE today are the leading forces on bringing the Desktop on Linux and KDE is 5 years ahead of what's on GNOME today. They are not in the need and not in the mood (from reading the kde-core-devel Mailinglist) to re-invent all sort of libraries or adapt poorly designed GNOME components in KDE. Havoc Pennington is arguing on named List that people should not care wether they use OpenOffice, Evolution, Mozilla and other components on whatever Desktop but the real point is that we on KDE have no real need to run these applications because we already have powerful counterparts for them. We understand that GNOME lacks serious Office suites and really nice integrated Webbrowsers but that's not our problem. See, they came to us and we not to them. You have and must understand that KDE has a wide acceptance even in the german government and most major Distributions offer KDE as default Desktop. We deliver the Desktop and the Tools for a wide area of people including real business and corporations. What do you have to offer on GNOME that we couldn't offer on KDE ? What business applications (that don't crash) can you offer for business ? See, we are not in the bad position after all. KDE since the version of 3.0 has a stable, documented and working framework and we offer a lot of applications for business today. Application development is a rapid process these days. People need half (if not a quarter) the time than you on GNOME need to develop applications. Not to mention the poor documentation of your libraries and the lack of poor programming manuals will make it take 5-6 times longer than normal because your developers need to spent more time finding out how things has to be done before they can do it for their own programs.
Think about this and this said, insulting me as person only shows your immaturity.
Why do you refer to me as a Gnome developer? I've always been a KDE supporter but the arrogant attitude from the KDE camp to the rest of the community caused me to speak up.
You have a way of making KDE look like the future dictator on the *nix desktop that don't obey standards just sets them. Should KDE make it as difficult as possible for other developers to compete? Even the commercial sector recognizes standards as a vital point to ensure competition and innovation why shouldn't we? Most of the things Havoc asked for only does not require KDE to make big changes and they would be a big win for the users and developers of DE independent apps.
Dude! get a grip.
So we can't say we have the best desktop... that KDE rocks? (and GNOME sucks :o)
If Havoc wants the changes so bad, he should just do them himself. if it's any good we'll use it.
you want standards????? drop GNOME and use KDE.
I agree
Youre free to cheer your team as much you want, but dont let the soccer mentality take over completely.
>If Havoc wants the changes so bad, he should just do them himself. if it's any good we'll use it.
You really believe that would be the case?
>you want standards????? drop GNOME and use KDE
Standards shouldnt require settling on one solution.
With great power comes great responsibility ;-) It's time KDE takes this into consideration.
You want standards. Drop Linux and use Windows.
Reading stuff like this impresses just how possible evil things like genocide are. This guy has turned a desktop environment into a characture of evil!! Imagine what people like him could do to enthic scape-goats!
I am a gnome user and I personally thing Havoc is a very sane and professional engineer. I like gnome2 much better than gnome1. I also like KDE. I think your insane. The poeple here trying to turn Redhat into Microsoft, or develop wild conspiracy theories are the one who are treuly sad. Get over yourselves, KDE and Gnome are on the same team!!
Seesh,
Ryan
> Get over yourselves, KDE and Gnome are on the same team!!
If this is the case then why are they working on 2 different things ?
I dont understand your question. If US and Britain are on the same team, why are they different countries?
This is definitely YXALAG.
Get a life buddie.
Wow. There's a lot of bitterness and hatred out there, and over something as silly as interoperability. In some ways, I'm glad that a lack of "_____" in gnome means people like this "KDE" user is well, a KDE user.
This arrogant GNOME attitude is really tiring. Why can't everyone get along?
More on GNOME arrogance: http://bugzilla.gnome.org/show_bug.cgi?id=95273
This one is funny too, same person. Dictator Havoc Pennington.
http://bugzilla.gnome.org/show_bug.cgi?id=90012
Do I sense jealousy here! He He Someone is very jealous of Havoc.
You will never be half the coder that he is, no matter how much FUD you spread about him.
You don't need to be half the coder Havoc is when you program with KDE/Qt! :-)
Havoc must be good, because he is the only one knowing all the *undocumented* GNOME libraries.
Wow, that Ali in that bug report looks like a total asshole! Flaming using Bugzilla? Incredible...
I'm that Ali who reported #90012 on b.g.o and I didn't flamed him. I got upset because of Havoc Pennington's behaviour of the way he replied to me and others (read the crossreferences and similar reports to that bug). He replied to others the way that gives the individual reporter to be the 'first one' who reportes such strange behaviours while he got multiple times the same report from various people. He then put up the reference that the OTHER programs where buggy because of not following the specs from freedesktop.org while the other people probably never heard of freedesktop.org in their whole life. Havoc Pennington invented that WMHINT, declared it as STANDARD, implemented it into MetaCity and made it incompatible to various other applications. Try running a new MetaCity and try running Xine, Mplayer, Phoenix (Mozilla) and various other and older applications in fullscreen, they don't work and this person has the nerve telling me to report it as bug to all the other applications while he was responsible on his own for implementing a new WMHINT which he has declared as new standard. Sorry but if you call me an total asshole then you should better investigate into the backgrounds (which I have told you now in form of an excerpt). I tried to talk to him normally but his hardheaded way of dealing with other people forced me to talk differently to him and as you can see a lot of other people where standing behind me and backed me for my reply (not nice after all I agree but it had to be said during that time).
I wonder why he can't get it through his thick skull that if he finds himself stepping on everyone's toes, then something is wrong with him. God! I won't be surprised if this guy has been kicked out of the millitary, 'cause he can't deal with authority. In short he can't deal with anybody that knows better than him.