KDE 3.2 Beta 2 was released last week for general testing and OSNews offers a preview of what is to expect from it early next year upon its release. The article mentions KDE's new features (faster loading times, Konqueror's Service Menus, Kontact, KPDF, Plastik theme etc), the problems that still plague it (cluttered KMenu and Konqueror menus, too many disorganized control center modules) and some constructive suggestions on how to get over the bloat without losing the functionality.
Dot Categories:
Comments
> The KDE fanclub changes their mind on styles every 3 weeks, but if you go back to just before 3.1 came out everyone thought Keramik was the coolest thing since sliced bread.
The difference is that not a large amount of people actually TRIED keramik before 3.1 came out, because, they needed uh, CVS or unstable releases. While, since Plastik started out as a seperate style, it was used by a lot of people before inclusion into CVS.
Keramik might look neat to look at, but it feels horrible to *use*.
How can Keramik "feel horrible to use"? While the taste for different styles may vary a style "feel horrible to use" is clearly a matter of your imagination.
> While the taste for different styles may vary a style "feel horrible to use" is clearly a matter of your imagination.
Uh, Keramik does indeed feel quite bad for daily usage __for a large amount of people__. The worst thing is it's bulkiness. Plastik doesn't have such problems.
Sure, Keramik (iyo) feels bad, and? I also feel quite bad when I am forced to hear (imo) bad music or watched an (imo) bad movie, wasted my time reading an (imo) bad book. It's still all a matter of taste. Why are you trying to prove something you should know you can't prove beyond your own self? Taste never has been universal.
Actually there were tarballs available and SuSE even shipped Keramik with 3.0 with I think SuSE 8.1.
Again, I mostly chalk this up as everyone getting excited about the new stuff; most normal users really don't care.
This will be a valid debate for 4.0, but changing the default style twice in a major release is simply a bad idea; like it or not there should be some continuity in a major release.
Well, actually they do. Transparent windows are now gone, brushed metal in the latest version. I believe that Eugenia mentioned these these changes some time ago.
> Can you imagine Mac changing the default look and feel with every minor release?
Apple actually has with OSX 10.0->10.1->10.2.. with great success too, since they've been making a more and more polished style with less "in your face" eyecandy each time.
The days of lickable buttons are fading, at least until 2005 when longhorn comes out :-)D
I use/rely on BOTH KWrite and Kate on a regular basis (ie about 18 hours/day) for exactly the reasons above - Eugenia just needs to upgrade to KDE and she would understand:) (btw the docking file listing in Kate 3.2 seems broken since it overwrites the editing screen and doesn't stay active although the tear-away feature is very nice).
However, the number of editors in the menu even struck me as too many (maybe Kommander should be taken out of editors, moved somewhere else) and yes, there seems little reason to have both KWrite and KEdit (I realize the differences - but they could be combined - I always use KWrite even on simple text config files).
I thought that first, too. But I found out that you have to click on the rectangle (the icon in the middle of the three really tiny icons) in the top right corner of the docking file list. Not very obvious...
Kate! Kwrite! KEdit! Kate is not representative of its function, Kwrite sounds like a wordprocessor, only Kedit has a name respresentative of its function. Even the names are confusing! Most users, including me only need one and that one is Kate (should be renamed to Kedit)! Kate has a very configurable suer interface and it can do just about everything Kwrite and Kedit can all in a single package. THERE IS NO NEED FOR THE REST,s top trying to pelase the last 5% who want this. This will cause nothing but ambiguity and a detremental effect on useabiltiy and space, if KDE IN THE OFFICIAL DISTRIBUTION can not decide on a single all purpose text editor, this is really a sad day for the project.
Really speed is great in Kate alsoa nd that is not an issue anymore, not for something like a text editor, Kate opens in less than a second here and hardware is getting far faster as is software so there is no need to nitpick on speed so much. By the end of 2004, Intel should havea 3.8 GHZ computer for example and most will have at least 1-2 GHZ.
The problem with your one editor is that from the limited perspective you're looking at it with it makes sense. Many people with much greater insight and perspective have looked at it and having all three makes sense. What exactly do you gain in taking away from others and why exactly is it a "sad day" when someone who ccouldn't answer dozens of questions the developers had to doesn't agree?
What is sad is restricting choices in ways that hurt others... especially when you don't even know who it hurts and how.
How do you know how much insight I have? Also Eric, I already mentioned why this is a bad idea in my original post. I am not proposing the immidieate removal of the rest, I am jsut proposing that development be focused on one and the others just remain as legacy until KDE 3.3 or 4 when KAte cand o everything they can and more, meaning that they can be removed.
As for restricting choice. This is not so, I only said that it shouldn't be in the main KDE distribution, these applications will still exist. So if the user feels Kate is inadequate they will just find a better one.
You are running into open doors. KEdit is no longer in the kdebase package and is still there only due to the missing bidi support in the other editors, and KWrite and Kate base on the same code so there's is no development time wasted.
So to be blunt, I wish you much more insight than you are currently showing.
Kate can't do bidi yet. After that KEdit might be dropped. I just had a terrible experience with it: replacing the same text in 2800+ took so much time that I killed it. In Kate it took a fraction of a second. Anyway, I'm happy with Kate and KWrite. :-)
Andras
> and most will have at least 1-2 GHZ
Spoken like someone with cash to burn.
"why would I want to zip a web page or use Cervicia with it, is beyond me"
Is she expectiong an answer? I want to zip a web page to save it to disk and view it later, and use cervisia to admin a page remotely. She doesn't, then criticizes.
"Konqueror is the Frankenstein of file managers, made of so many Kparts that the end result is just not good"
Then she's misunderstanding the nature of konqueror: it's made to be a Frankenstein and to hold as many kparts as one wants to throw at it.
"Another thing I dislike is that "settings" menu that most KDE apps have, where they list 3-4 different "Configure" options in addition to the "Configure the application" option"
This comes from above: she misunderstands the nature of KDE and reusable parts. That's intentional, the applications are "frankensteinized" and every part has its own life, get used to it.
"there are too many apps shipping with KDE"
And others complain of too many apps for "alternative platforms". Come on, most of them are optional and the distributions pack them separately.
On the other side:
I agree on the point of making the color of the icon text automatic, if someone can hack it (I can't think how to do it, sounds a bit hard).
"expanding a few *pixels* the space between words on the application menus"
I'd advocate a measure on ems, the current about 1em here can be probably improved by the about 4em of gnome (in her example). But I don't mind too much; it's "good enough for me".
I couldn't make any sense of her criticism of kcontrol, so I'll skip that commentary.
I think she clearly has some localized problems. Her menus are far larger than mine and I install all of KDE from CVS. I think her biggest problem is that she installed everything that came with her distro and KDE just picked them up and added them to the menu.
I also think that Konqueror is not to blame for services adding to the service menu where it is inappropriate to do so.
KControl might benefit from an "Advanced Mode" in the future which activates modules that are initially hidden. We know that all users will eventually turn this on, but it will look more "accessible" to the user at first I think.
Over all the criticisms were not too bad but they were perhaps slightly misdirected or at least the real problem was not fully understood.
I fully agree with an basic and advanced mode for configurations.
As far as I understand, Cervisia is only of interest if you use cvs. I think cervisia is a kind of option that can't be categorized as simple vs advanced - rather it is a specialized option. It does not fit on a simple - advanced scale.
Cervisia is part of the kdesdk package (That's K Desktop Evironment Software Development Kit for you) so if your distribution installed it for you complain to your distribution that you don't intend to develop using KDE.
Besides I don't know how she managed to get them, but I don't have Cervisia's context menu entries while browsing through the web at all (and I do have kdesdk with cervisia installed).
Thats what you get if you are browsing the web in Filemanager mode of the konqui.
Could Kcontrol be laid out like the Mozilla Firebird setup screen. For those who haven't seen it it's like Outlook. The top 5-6 items could deal with basic things (Desktop Appearance, Application Appearance, Sound, Fonts, Personalisation, Localisation) and the last item could be "Advanced" which brings you to a tree where you can pick the appropriate values and edit them.
As regards context menus, they are getting a bit out of hand (and I've used KDE since 0.99). For one thing, all the XML-Gui extensions for K3B etc. should be put in a sub-menu (e.g. "Actions") just under "Open with...". I'm using 3.1.4 and it's crazy. Also you've got to wonder about some of the options added to these menus. The desktop _context_ menu should list things to deal with the _Desktop_ but it's got logout/lock buttons (they're on the shagging kicker!) open terminal and run command. None of these have anything to do with the desktop. And for those who bemoan the lack of use, I always click on kicker for loggin out, use Alt-F2 for running stuff, and have a terminal icon on my kicker, all of which is faster and easier than finding a free piece of desktop, right-clicking on it, and selecting the appropriate option in the menu.
> As regards context menus, they are getting a bit out of hand (and I've used KDE since 0.99). For one thing, all the XML-Gui extensions for K3B etc. should be put in a sub-menu (e.g. "Actions") just under "Open with...". I'm using 3.1.4 and it's crazy
Have you tried 3.2b2? The context menu is a lot better, but not perfect yet.
"The desktop _context_ menu... it's got logout/lock buttons..."
Hey, calm down! I use them all the time if I am using the mouse (with keyboard I have logout set to ctrl-alt-d). And each time I am in windows (not too often) I attempt to logout from a right mouse button click; annoying. You do it differently; let other people be.
Damn right. Also, most other WMs have logout on the desktop, who is KDE to break a fine tradition? ;-)
Of course I want to be able to lock my desktop, why shouldn't that be in the desktop context menu?
One of the great things about KDE is it's configurability, the possibility to let it do (almost) *exactly* what I want.
I agree with Eugenia on some things (for example I also find Keramik extremely ugly) but unlike her, I admit that it's a personal perference and not some kind of law written in stone.
There is no best setting for everybody. The defaults should try to satisfy as many people as possible, yes, but please NEVER EVER remove any configuration possibilities. I gladly invest a couple of seconds to find an option which will save me frustration and time FOR MONTHS. - You only configure a feature once, but you use it a lot longer.
Also, Eugenia is contradicting herself: She sais that computer illiterate users never change the defaults (which is true) but then she sais that the configuration is too complicated for them - If they never change the defaults anyway, they WILL NEVER SEE THE CONFIGURATION PANEL!!!
Eugenia has raised some very valid points but struck a nerve whilst doing it. I think we can summarise her by saying that KDE has three fundamental flaws: Look&Feel, Usability, and Compatibility.
Now I am sure we would all be very happy if further improvements were made in the L&F and Compatibility areas, and Eugenia's comments regarding menu spacing and toolbar separation is the type of stuff that makes Gnome look so much better. But please be careful when messing with the usability demon.
I do wonder whether the Gnome/MacOSX approach is truly better. I certainly don't prefer it, and nor do many others. But, being an advanced user it is not immediately obvious why KDEs way of doing things may not be perfect for everyone.
Still, good UI design comes from following a cycle of behavioural analysis, innovation, and product refinement, and I can think of many areas that could be substantially improved without making huge changes. But Eugenia's idea of splitting program configuration in two parts seems more revolutionary, than evolutionary. What proof is there that this is a good idea. I worked on a large security product that took this approach and we had a hard time making it work, ending up with a product that was inconsistent and even harder to use.
I am sure there is a perfect compromise in there somewhere, but perhaps it will only be found through careful, managed evolution.
This review was horrible, even for Eugenia. Anyone remember her review of Red Hat 8? If not, go look it up; if you read it and find yourself nodding your head in agreement, please switch to a more user-friendly operating system; you're not cut out for the world of Linux. In fact, if you agree with Eugenia's opinion about Red Hat's X configuration system, including its inability to deal with Eugenia's oddball monitor and graphics card, go buy a cabin in a remote location and swear off all forms of high technology; you can't handle any of it.
The entire review is an editorial, and a poorly-written one at that. And yes, as usual, she's managed to contradict herself. She complains about things that, by her own admission, most people will never see. And quite frankly, the bit about using Plastik over Keramik is purely opinion; I use Keramik, and find it to be nice-looking. The default colorscheme, on the other hand, is horrible in my opinion, as is the window decoration. Scrollbars could use a bit of work, too. But that's just my opinion. :D
I agree that KDE shouldn't make the same mistake as GNOME. I don't recall the name of GNOME's default CD ripper, but the last time I checked, it didn't even offer an option for bitrate. Is the average user REALLY going to get confused if they get to choose a bitrate?
>I agree that KDE shouldn't make the same mistake as GNOME. I don't recall the
>name of GNOME's default CD ripper, but the last time I checked, it didn't even
>offer an option for bitrate. Is the average user REALLY going to get confused if
>they get to choose a bitrate?
Sound Juicer is just not finished yet. Ross Burton says he plans to add that feature, but just hasn't yet. but to answer your question, yes, it will confuse the average user unless it's couched as a quality scale (low, medium, high etc) possibly with the bitrate being show for advanced users, but possibly not. 'cause seriously, who cares?
>yes, it will confuse the average user unless it's couched as a quality scale (low, medium, high etc)
That's a terrible idea, IMHO. What about people who want a semi-predictable filesize?
>possibly with the bitrate being show for advanced users,
>but possibly not. 'cause seriously, who cares?
That's an awful lot of assumption.
No matter if what she says is good or not, her review is useless. I can't trust on somemone that hasn't even checked the facts that she claims:
" I don't need four text editors in the same submenu (Kate, KWrite, Kedit and Kommander-something)"
Since when is "kommander editor" a _text_ editor? She didn't even bother clicking on anything. That says all.
Maybe Kommander should add a GenericName to its desktop file to specify what it is. It may not be a text editor but it didn't tell me what it IS instead then. Even the "About" box did not say anything other than "Kommander Editor".
Editor for what?
True. I didn't notice that fact. One can't easily guess what it does, unless he has used Qt Designer before.
Right. Any suggestion? GUI Script Builder??
Andras
Yes that sounds like a good name to me.
Fab
Kommander has reached a level where a moderately capable user can do a fair amount with it, but unfortunately it lacks some polish. With some work it could become a tool that could enable less experienced users to do much more. So we need to address it, but we also need more developers as Marc has not been able to give it much attention and I'm debating taking the lead coding it. (as if I had time) Anyway I loaded things up and noted it showed up as an editor. I never really thought it belonged there and don't know who put it there. It really should go under development tools.
I would also agree with the critisism of the review. Apparently there are too many programs because she can't open them all and tinker with them a few minutes in the short amount of time she allots to review. After the recent review from a BSD user who did take the time I'm forced to agree that the packages should not be made to fit into her reviewing schedule, but it should be the other way around. No wonder she's obsessed with visuals.
May a kind of User-View options practical?
like:
novice-> less less and simple options
medium-> opitions like now
hardcore-> more (every) options
Sure, and then produce 3 versions of the documentation (each with different texts and screenshots), and 3 versions of all KDE web sites (because people would otherwise wonder why the screenshots look so different from their desktop), and different user groups (because a hardcore user can't help a novice - the hardcode user does not know which option the novice has)....
That has been brought up on kde-usability many times and it has been turned down every time. It will just not work because different users are good (and bad) at different things.
Well, maybe grouping some of the modules and adding an 'advanced' button could do the trick (it wouldn't reduce functionality, but should look cleaner) eg:
One 'Appearance' module, containig 'Fonts', 'Windows Deco', 'Style', 'Colors', 'Background', the first tab would be used to select a style (Plastik for example consists of style, colors and window deco, so a unified setting would work) a background (image or color) and one base font (eg. default all fonts to Vera). Klicking the 'advanced' button would ungray tabs to further customize the settings.
I think I'll do a mockup and post it (on KDE Bugs/ KDE Look) soon...
and what about 'advanced...' buttons .. impossible too ?
Why then on kprinter do we have a expand and collapse section and a properties button. If we need to get to the options we can, but for the simple stuff - keep it simple.
Somebody (maybe I will) should compile a list of every option in kcontrol. The list should be a flat list without hierarchy. Options should be listed as those that are basic and advanced - voting and marking could possibly done on a wiki style page. The basic options (probably 1/5 to 1/3 of all) should be put into refactored dialog boxes. The remaining could either have expand/collapse buttons, an advanced tab, could have their own dialog section, or could even have a simple section similar to the kprinter driver dialog (which is very minimalist).
I'm sure the developers have voted against it. It is hard and who wants to decide what is basic and what is advanced. At somepoint it will need to be done. And if I, who knows relatively little about good C++ programming, have to grab CVS and submit a patched kcontrol - then maybe I should. Or maybe the people who developed it originally and understand its internals and feel ownership should do it and get a much better result. Either way, by the next KDE release 3.3 or 4.0 - somebody will have done it - whether the mailing list agrees or not - it will happen (even if it comes down to me getting off my soapbox and doing something about it myself).
I mostly agree with Eugenia whe she says kcontrol is too cluttered, and a simpler control panel should be used instead.
But i think a gconf-editor/regedit like solution would be horrible and an almost inusable annoing program.
One way could be inserting the basic/advanced button in kcontrol, or shouldnt' be instead better if the settings:/ kio-slave would show fewer and most common items like background or color settings and leave kcontrol as is? The kio slave would be the easy panel and kcontrol the utility fon fine-tuning the desktop and no regedit-like application would be needed, because in some things kcontrol likes regedit a lot(like a tree organization or the search function) but doesn't have that annoying behaviour key/value key/value key/value.
And, going a bit off-topic, another important fact is that the kio slave could show either shortcuts to kde configuration modules and proprietary configuration modules added by the distribution.
Look for example to Mandrake, there is the Kde control center and the Mandrake control center, why two control centers in one single desktop? this is absurd, but the Mandrake configuration tools are written in gtk so at the moment is not possible to integrate them in the kde control center(like Suse did), but the kio slave starts every module in a new window,so this should not be a problem.
In fact these tools are yet integrated in the Gnome control center, as entries in their settings vfs, giving more consistency to the desktop.
This message is posted from Konqueror in Kde3.2_beta2, and I think KDE3.2 is the best desktop environment I've seen. But, I have to say I agree with many of Eugenia's points. My pet peeves are:
- The context menu in konqueror is nasty
- The control centre is a nightmare.
- Default settings are generally horrible. Keramik is ugly and too large. Toolbar layouts are strange and cluttered. The close/maximise/minimise buttons are right next to each other. Why not enable icons on buttons by default?
I'm going to go see if usability.kde.org needs some volunteers.
Icons on buttons are not on by default because not everyone has a computer as fast as yours. It costs to add those icons.
Note that the only way to reduce the context menu significantly is to remove functionality that people want and use. We did try removing entries during development and some features became unusable or the number of complaints far outweighed the number of people supporting it. Cascading the menu is no real option either as it really only increases the number of options while making them harder to access. If you really don't like it, you can configure it. It uses and XML file to define what's in it. Remove the features you don't want.
I didn't realise button icons were that heavyweight. It's just that they decrease the time to figure out which button does what - it's immediately obvious that green=yes, red=no in general. It seems a good feature for newbies to have.
I will go edit my menu XML. It would be nice to have a menu type selector or something though. Anyway, I'm going to quit being a whinging user now.
Those colour associations arn't valid in all cultures!
not to mention colour blind people and monochrome screens (still present in the embeded world) ;)