Three reviews recently joined the list of KDE 3.2 reviews with most having an emphasis on the applications being new in KDE 3.2: Francesc tried it on Gentoo and states in this blog "This release is the best KDE I've ever tried". Pycs writes about his first impressions and likes it too. Finally, gooeylinux.org published a rather controversial article about the good and bad sides of KDE 3.2 Beta 2.
Dot Categories:
Comments
And the biggest point really is that HIG mention, just try to install and customize a dozen machines at the same time and you see why that makes perfect sense.
And yes, I really miss some of the other features that are mentioned on that article too.
Even when I'm using kde-3.1.94 as the only desktop installed... ;-)
Perhaps some points with many more proven wrong (see OSnews comments) points.
He's right about Konqueror's rendering in 3.2 as well.
Yes, it feels like the merging stopped at half the way. But hopefully the emerged regressions are fixed during 3.2.x lifetime. For 3.3/4.0 it will be interesting to see if Webcore, being a faster developed and better maintained fork, will be picked up as main stream by the KDE people.
I agree, I'm rather disappointed by Konqueror's rendering. I was hoping it would improve more, especially now with Mac's Webcore. Bug #31417, where BBC News and CS Monitor sites have text going over other text and images is probably the most annoying, its one I see fairly often. I like browsing with Konqueror since it opens so quickly and its easier to open and save stuff (more integrated with environment) then Firebird.
One of my favorite new features in KDE is FSViewPart. I've found it to be quite useful.
Some called this review professional, I have to disagree for the following reasons:
He supplies us with unimportant data like his hardware configuration. But forgets to mention the distribution he is using and how he installed this beta2 on this system.
His installation is seriously f***ed up, since simple things like the desktop hiding button, and pdf's do not work.
He hasn't got a clue what 'memory usage' means, but blames this 99% number on KDE.
He contradicts himself (too many buttons, but one should be added for smb-browsing, just because he browses smb-shares)
He didn't do any research (yes plastik is the default, and yes there is a kde front-end to mplayer).
He thinks everything is bloated, unintuitive and inconsistant, but he provides very little evidence.
This is not a review, this is a rant.
> This is not a review
Or a terrible review.
It makes you wonder what he's up to...
Yes, it is a terrible review.
Read what he said about: NoAtun/Kaboodle.
AFAIK they ARE a front end for Xine. At least anything I installed for Xine worked for/with them. Including QT movies with sound. I thought that it was great except that building Xine from source wasn't a whole lot of fun. :-)
I can only conclude that he dosen't understand that you have to install the libraries and/or Windws DLLs in Xine for the respective audio/video format to work.
ARGH!!
--
JRT
They are not a front end for Xine in the fact that they'll build and play some media files without Xine. However if you've got Xine there's a Xine pluging for arts which allows more types of media to be opened.
Thank you for your technical correction.
IIUC, if you install a library or Windows DLL for Xine then: NoAtun/Kaboodle will use it.
Isn't that what I already said?
So, just exactly what is your point -- you obviously didn't understand mine.
Perhaps in your rush to find something wrong with what I said, you failed to read the "rather controversial article" which says:
<<
NoAtun/Kaboodle - These lack features. The interface is very clean, but that's because it can barely do anything. It wouldn't play most of my video files, that other players have no trouble with. This is unacceptable, as there is mplayer, and lots of KDE front ends already around. They should simply take one, stabilize it, and use that. Or if they desired, use a XINE frontend. Both would be much better.
>>
Now do you understand *my* point that the reviewer is dead wrong because KDE already HAS a frontend for XINE. It is nonsense to say that this isn't the case because there are one or more additional layers of software envolved or that N/K also act as a frontend for other libararies. If the reviewer had installed XINE and the necessary libraries and Windows DLLs then N/K would have been able to play the files which he complains about.
--
JRT
"He didn't do any research (yes plastik is the default, and yes there is a kde front-end to mplayer)."
No plastik is NOT the default.
Oops, well I don't care about theme's anyway.
> And the biggest point really is that HIG mention, just try to install and
> customize a dozen machines at the same time and you see why that makes
> perfect sense.
what exactly does a HIG have to do with customizing a number of machines at the same time?
Obviously didn't mean only that, that just happen to be one of the accidental happenings what happen when all (specially desktop) config options are spreaded on multiple places... :-)
I find it always remarkable that people manage to say that a reviewer has it "wrong". A reviewer usually expresses his opinion, and opinions can't be wrong. You can agree or disagree with an opinion though.
Take for example his comment about memory usage. He is under the impression that KDE uses 99% of his memory probably because some tool (top? ksysguard?) reports this. Does KDE use an unreasonable amount of memory? Not really, but memory reporting tools under Linux are so crappy (or, depending on your viewpoint, the way in which Linux uses memory is so complicated) that meaningful numbers are impossible to collect by mere mortals.
Something else what a lot of people don't seem to get is that a reviewer writes reviews. A reviewer isn't usually part of your project and isn't writing reviews as part of an elaborate bug-report. It is a report about the state of affairs as the reviewer sees it. Compare it with these radar-speed-controls the police does on highways which, unlike popular believe among some, are not there to calibrate the speedometer in your car either. You should have had the right speed before you got there.
Then this review has a remarkable amount of wrong facts expressed as opinion. Sorry, declaring everything as opinion doesn't make this guy less clueless.
IMHO, you can't blame a reviewer for not being a technical genius and getting some facts wrong. Every user opinion is important, whether its technically correct or not - except KDE should only be used by hackers. Someone very familiar with KDE (and Linux) will probably (unconciously) overlook the many minor (and not so minor) glitches that still trouble KDE and other projects.
Recently I have seen many times on help forums people believing all their memory and even more was used because they used ksysguard without knowing what it displays. And for the beginner, there is no warning that ksysguard is not that easy to figure out.
That "rather controversial article" was written by a 16 year old child. For a kid, it's a decent review. But I wouldn't place too much stock in it. Use KDE 3.2 yourself and form your own opinion.
Well well well. Seems to me he's whining a bit too much.
A lot of buttons and stuff that "clutter" the interface?
I see this same comment averywere when you read about Linux and KDE.
I think it's up to the reseller/distribution/whatever to
do the final twinkles, just look at Lycoris etc.
The basic KDE should show off all functionality.
> A lot of buttons and stuff that "clutter" the interface?
> I see this same comment averywere when you read about Linux and KDE.
Yes, it seems to be a "hip" thing to say.
The fact that he requests SMB-settings to be moved to the default interface (cluttering up the interface even more) just shows that he's just parrotting whatever he hears and is in fact so clueless that he doesn't even realize how much he is contradicting himself. (Whoever runs a 2GHz machine with 256MB RAM is obviously clueless, I'm surprised that KDE seems to be still running quite fast on his machine, but that's another topic.)
I want a desktop that I can use efficiently over years and that's exactly what KDE is. You can configure it (almost) exactly the way you want.
I do not want some piece of junk that looks great in the first 30 minutes of usage but then starts to slow you down with useless animations. Like MacOSX for example. That any usability expert could find MacOSX good remains a mystery to me. In the release I tried (IIRC 10.1) there is no way to turn off minimizing animations (I could choose between two different kinds of animations which added a lot of irony to the whole matter) which drove me nuts after an hour of usage and is just one of many examples how completely and utterly useless MacOSX is for day-to-day work. Yes, the icons look great at first, but at the end of the day they are completely useless and don't do what icons are supposed to do: Differenciate applications on a small screen-area. KDE-classic icons are much more efficient. (= *much* smaller, yet still easier to tell apart from each other) Unfortunately the KDE-keramic icons are also drifting to "consistency" which is the same as "useless" when it comes to icons. (The job of icons is to look DIFFERENT, not to look similar. If all icons are blue, it might look great, but it becomes useless) Another bad MacOSX example is the pretty useless dock which takes up much too much screen estate and becomes useless as soon as you have several windows with the same icon (Text may not be as sexy as huge graphic icons, but at least it does the job). The lack of multiple desktops and real 3-button support just completes the picture.
Anyway, I have the strong impression that the review-community is just a bunch of guys parroting "general truths" (like: "If it's from Apple, it's the most usable thing ever created". Take the horrible "puck"-mouse for example. When first released, nobody dared to critizise it, it's from Apple, the usability-gods. It took many years until the reviewers dared to critzise it. And their newer mice aren't much better either, it's ergonomy and usability sacrificed for good looks.).
However, my point:
It doesn't matter what the KDE-team does. The "general truth" is that KDE is bloated, inconsistent and confusing. Reviewers will repeat that general truth regardless of KDE's actual look and feel, just like they will praise anything from Apple without any concern about the real usability.
So please don't take such reviewers seriously, especially when they contradict themselves within a single paragraph.
Some additional comment:
Take the "KDE is like Windows"-myth for example. For years it was repeated that KDE is just "like Windows" and therefore evil and that real men would use GNOME.
After GNOME has started to dumb down everything and copy some horrible Windows-features (like the Registry -eeech), all of the sudden KDE is no longer "like Windows" any more. Now all of the sudden it's too complicated and confusing and GNOME is the better Windows.
Did KDE really change in that regard? Not really, it always had loads of options, yet the public image turned 180° WITHOUT KDE ITSELF CHANGING!!
There is one area in which KDE is really lacking and that is marketing.
Apple has people creating the public image of great usability.
Microsoft has people creating the public image of universability, compatibility and low TCO.
GNOME has people creating the public image of easyness.
KDE has nothing comparable and that's why the public image of KDE is driven by the marketing guys of other organizations.
Again: IT DOES NOT MATTER wether KDE follows usabilty-guidelines. It's irrelevant in influencing the public image. What DOES matter is that such guidelines and some usability team exist and that somebody supplies a constant flow of press-releases about KDE's usability and how great it all is. After some time of constant bombardment with usability press-releases, reviewers would start parroting it and a new image is born.
> copy some horrible Windows-features (like the Registry -eeech)
What's wrong with the Registry/gconf? The only argument I've ever heard is that the Registry can be irreversibly corrupted, which, while valid, doesn't apply to gconf because it uses XML instead of a binary format.
> What's wrong with the Registry/gconf?
Let's see:
1) It complicates everything by introducing several completely inconsistent ways to do configuration.
2) It prohibits easy upgrading. I still use ~/.alias, ~/.profile and many other files back from 1997, but I wouldn't want to try to import a Windows98 registry on WindowsXP, I wouldn't even dare to do it between servicepacks. In general sharing configuration among Windows machines is pretty much unheard of (it's either a complete diskimage or nothing) while it's very common in the Unix world.
3) It's not commented and it doesn't support comments. Just by reading /etc/httpd/httpd.conf you get a pretty good idea and should be able to do a usual and ordinary configuration without too much problems. You can easily go back to the things you had in place before by *commenting out*. It may be a trivial and "unworthy" reason, but the "
In reference to (1) and (4). GConf uses XML, not binary, so its nothing like the registry. Its no different than a tree of text files in /etc, except its all in the same format and has a standard API to read/write it. And posting XML to newsgroups works fine, since browsers usually do not try to interpret it.
XML can still contain binary parts (dunno if GConf supports that atm though), and is far less readable than the .ini inspired clear text style KDE uses.
GConf supports aribtrary back-ends. They use XML simply because they didn't want to invent yet another format. Using XML allows for the storage of very complex configurations, and allows people to use widely available, standard XML tools to manipulate them.
And yes, XML supports binary parts, but in theory so does .ini. Just uuencode some binary data and stick it in the file as a string. In some special cases, you might want to do just something like this. There would be a problem if gconf used binary data on a regular basis, but it doesn't.
1) No, gconf provides one way to do configuration. ~/.blah files provide inconsistant ways to do configuration.
2) Part of the gconf policy is that apps should always be backwards compatible with gconf settings of previous versions. I don't know how well that's followed, but it's not any worse than the format changes in ~/.blah files. As for sharing configuration, you can access the settings on one gconfd server from any number of network clients.
3) Yes, it does support comments. gconf has a schema for each key describing the key. Open gconf-editor and select a key, and a description appears. No, you can't comment stuff out, but you can revert keys to their default settings.
4) So, instead of telling the mailing list "my ~/.blah file says '[foo]=bar'", you can just say "foo is set to bar". It's just a difference in semantics. It may not be quite suitable for complexly-configured apps like Apache (although 99% of Apache "help" requests are configuration problems caused by it's own unique config file format), but it's quite usable for everything one would want to do in a GUI desktop environment.
Another benefit of gconf is that it makes it much easier to write apps. You don't have to invent your own format and code your own parser, everything is done for you. You can also do stuff like to register to be notified of changes to a certain key, so your app can react instantly to an external configuration change.
"Did KDE really change in that regard? Not really, it always had loads of options, yet the public image turned 180° WITHOUT KDE ITSELF CHANGING!!"
Yes, but others have changed. KDE 1 was a breakthrough for Unix usability and KDE 2 was an awesome upgrade. But since then, it has mostly been configurability and features plugged on top of it, while others didn't stand still.
You are lying to yourself if you really think, that everyone who doesn't enjoy using KDE anymore would just be following a trend or beeing manipulated by PR.
As the KDE project gets seemingly more and more complex people from outside who don't care to inform themselves beforehand just seem to become more and more misinformed but still state being capable of judging everything from the top to the bottom. And KDE is certainly not the only project which "suffers" on that.
"Did KDE really change in that regard? Not really, it always had loads of options, yet the public image turned 180° WITHOUT KDE ITSELF CHANGING!!"
Haha! Thanks for bringing this salient point to the forefront of my consciousness! Yesterday KDE was bad because it was too much like Windows. Today it's bad because it isn't enough like Windows. It reminds me of the Qt license issue. No matter how free Trolltech made Qt, people bitched that it wasn't free enough. I've come to the conclusion that a sizable and vocal segment of the community hates KDE for the mere sake of hating KDE.
Exactly. It seems like most reviews of desktop environments are written after a quick look round - click a few buttons, check out some apps and snazzy effects, then write the review.
Saying things like "it takes up 99% of my memory" shows the gooeylinux guy doesn't really know what he's talking about. He does have a few points, though. Konqueror's web page rendering is pretty bad, the default sounds are terrible and 'consistency' of icon design is not such a good idea if it means it's harder to differentiate icons.
> Konqueror's web page rendering is pretty bad
Yes, I also still need Mozilla from time to time, but in general it's pretty OK for me.
But I still remember the times I used KFM ;-)
Actually the feature KFM and Konqueror have and no other browser (I have tried) has is that it reopens all browser windows on login. On the right desktops with the right geometry and now also with the right tabs. No other browser can do that. (Opera can restore *one* browser window AFAIK and it is completely unaware of multiple desktops) So with Konqueror you no longer need temporary bookmarks, for example on a web-forum you can just leave the current page open and continue at the next day.
Now what is really needed is a press-release selling this feature as great and completely new feature in KDE 3.2. Nobody will care that it already was in KDE1.0...
This is the feature which made me use KFM...
> the default sounds are terrible
Yes, I can agree 100% on that. However you rarely read about that in reviews - it really seems that reviewers are somehow ashamed to write something new or original. I really wouldn't mind hearing constructive critizism but this mindless parroting just makes me sick.
Oh my God. Those Mac one-button mouses. Oh dear.
I think you are unfairly dismissing an entire body of claims just because a lot of the people making them are not terribly informed. I am a full-time KDE user since 2.x, and I agree to a degree with some of the things that people are saying. I don't believe that KDE is unusable because of its UI complexity, but I definately believe its hurting my efficiency.
The overall problem with KDE is that there are just too many extraneous buttons and menu items. Its not just confusing, it hurts efficiency in the general case, and is unasthetic. Let me give you a rundown of the default 3.1.4 Konqueror:
- The 'up' button really does not make a whole lot of sense in the context of web-pages.
- "Print" and "Print Frame" are inherently slow operations. Saving a fraction of a second by putting it in the toolbar instead of making people go to the main menu isn't really an advantage.
- Same thing with the 'Find' icon. Find is a process that requires going to the keyboard and typing in additional data. Thus, the time saved having it in the menu is very minimal compared to the overall time taken by the operation. Plus, since your hands need to be on the keyboard anyway, CTRL-F is faster.
- The security action doesn't even make sense. Its not really a command, but a way to get info about the crypto status of the page. It should be an indicator in the status bar, not an icon.
- Increase/Decrease font sizes is probably not used enough to warrent being on the main toolbar.
So the complaints *are* often legitimate, just poorly explained. But it seems clear to me that it is these sorts of simplifications that many of KDE's detractors are looking for. I've posted screenshots of my modified Konqueror on OSNews, for example, and people generally reacted quite favorably.
http://www.prism.gatech.edu/~gtg990h/
I think "reviewers" should go into this sort of detail about exactly what about KDE they find cumbersome, but in my experience, that's a sure way to get certain KDE folks up in arms about how they *absolutely need* that feature.
I'm of the following mindset. Apple/Microsoft/Sun's HCI research shows simpler, more streamlined interface is more favored by most users. Now, you can debate about the degree here, about how simple is too simple, and how simple is simple-enough, but its kind of hard to argue the basic conclusion. Emprically, some of the UIs widely hailed as the best ever (MacOS Classic, OS X, NeXTStep, Amiga) were on the simpler end of the scale. Thus, I think it is benificial to have a more simple UI be the stock configuration. It doesn't have to be ascetically minimal like GNOME. Certainly no features need to be removed or even hidden --- things just need to be migrated from the toolbars and context menus to the main menu. You'd be surprised how minimal changes in the *visible* interface can give people a totally different impression of the UI. And since KDE is supremely configurable, people could always add any oft-used shortcuts back to their configuration.
One key problem here is that users who like a UI with lots of buttons and menu items could have to go to a lot of trouble to add all the things they want back to the stock configuration. Unlike other UIs, I think KDE's technology can result in a pretty nice solution for people of both tastes. From what I've seen, XML-GUI could probably be used to allow people to save their toolbar/menu changes, and even share them with others. If people could trade UI "themes" like they trade KDE Styles, then the "minimal or maximal" debate could become moot for most KDE users, just like the default style debate is today.
I've been playing a bit with XML-GUI to do just that, and it seems entirely doable. Wrap things up in a nice GUI (along with a menu editor to go along with the toolbar editor) and the thing could be quite usable.
I agree, the exchange or at least offering of different levels of amount of toolbar entries is already fully doable and should be possible to be include as a eg. "GNOME" choice in kpersonalizer after KDE 3.2.
> - The 'up' button really does not make a whole lot of sense in the context of web-pages.
I use it all the time to get to the homepage. It's works always, is consistent (= works on EVERY page) and fast (when you are scrolled down somewhere on a page there is no link to the homepage in sight. Try to click the button and then move the mouse down, the menu pops up instantly (no delay) and you can move to any subdirectory.
> So the complaints *are* often legitimate, just poorly explained.
No, I think these complaints are pretty clear: Some people think that *all* users should be forced to use their personal preferences.
> - Same thing with the 'Find' icon. Find is a process that requires going to the keyboard and typing in additional data. Thus, the time saved having it in the menu is very minimal compared to the overall time taken by the operation. Plus, since your hands need to be on the keyboard anyway, CTRL-F is faster.
> - The security action doesn't even make sense. Its not really a command, but a way to get info about the crypto status of the page. It should be an indicator in the status bar, not an icon.
I agree here.
> - Increase/Decrease font sizes is probably not used enough to warrent being on the main toolbar.
I use it all the time.
> I've posted screenshots of my modified Konqueror on OSNews, for example, and people generally reacted quite favorably.
Isn't the fact that you could do all this without touching a single line of code proof enough for the superiority of being able to configure almost everything?
What is better? To show the user as many options as possible/reasonable and let him remove those he doesn't like (like you did) or remove everything except the bare minimum and let users add their options themselves? Your browser may look nice, but without bookmark-bar and menubar (how are users supposed to handle bookmarks? Or do you consider bookmarks redundant too?) I doubt it would be popular among users because adding everything that even Netscape 3.0 could do (ehrm, like bookmarks) wouldn't be so easy. Many users would dismiss Konqueror because they would think it can't handle bookmarks.
In general I'd say that it's easier to remove some unwanted toolbar buttons than it is to try all buttons yourself and add those you really need. You can remove icons in less than 5 minutes, but looking for features/icons and putting them in the right place takes a much, much longer time.
> I think "reviewers" should go into this sort of detail about exactly what about KDE they find cumbersome, but in my experience, that's a sure way to get certain KDE folks up in arms about how they *absolutely need* that feature.
You seem to *absolutely need* the feature to remove the menubar, a feature I never use. But I realize that it's just my personal preference and even though it's not me who would be hurt when that feature goes, I would argue for keeping it.
> Certainly no features need to be removed or even hidden --- things just need to be migrated from the toolbars and context menus to the main menu.
I could certainly live with that. But I would have doubts that a lot of features become pretty much unused and unknown.
If KPersonalizer would offer an "streamlined" versus "fully-featured" slider, I think that would be a very good solution.
> > - The 'up' button really does not make a whole lot of sense in the context of web-pages.
> I use it all the time to get to the homepage. It's works always, is consistent (= works on EVERY page)
You must be living in the web of 1997, because in the web of today, the 'up' button works in less than half of most web pages, and even less in most popular pages, which are oft-dynamicly generated and have weird physical layouts.
Web2004 != Heirarchical
Personally I still use it very frequently, even if it's only a mean to access the actual server URL without having to use the keyboard or cumbersomely selecting the URL part I want to toss away for then selecting "cut" from right click menu and pressing the "enter" button (yet another thing everyone seems to be eager to toss away).
The only way I see solving this is extending the already existing possibility of offering different schemes of predefined "defaults" in kpersonalizer (which runs after the first login) respectively from which the user can choose what he prefers.
Don't make the mistake of thinking Konqueror is a web browswer. It is not. Instead it is a network transparent universal browser. The up arrow is EXTREMELY useful for browsing the local filesystem and ftp sites, to give just two examples. It would be a usability mistake to make the up arrow disappear everytime you hit http...
The review complained that Konqueror didn't have any outstanding features. Well the "universal browser" is something most web browsers don't have. Many people have bragged this feature up big time, but for some reason the reviewers of the world just aren't listening.
I must be too, because I use the up arrow all the time. It's just so damn useful. Much of the time I use it to find stuff on a web site that isn't quite as organized as I would wish. I use it for exploring. For example, if I'm looking at a path like this: ../foo/i686/stuff, I might use the up arrow to try to see if there's an ../foo/i586/stuff directory.
Please show me a common webpage in which I can't go "up" to the homepage.
For dot.kde.org it works, I can use the "up" button from anywhere on the site to go back to http://dot.kde.org
In some very rare cases in which sites use different subdomains it doesn't work (most notably Slashdot) but those are exceptions, not the rule.
No, I think these complaints are pretty clear: Some people think that *all* users should be forced to use their personal preferences.
---------
A default setup implies that somebody is going to not like elements of the configuration. There is no way around that. You like having icons right at hand. I like them tucked away, because I prefer to use keyboard shortcuts. Currently, KDE is set up to please you and displease me. If KDE UI were more streamlined, it would displease you and please me. The question is to make the UI such that it displeases the fewest number of people. I would argue that KDE's current UI is not set up that way.
Isn't the fact that you could do all this without touching a single line of code proof enough for the superiority of being able to configure almost everything?
--------
I'm not against leaving in the configurability. I'm fully support leaving in configurability, even if that brings some complexity. I think that when the GNOME people say that KDE is to configurable, they're full of crap. What I'm against, however, is having the defaults be complex, when HCI research shows that most users like things to be simpler.
I could certainly live with that. But I would have doubts that a lot of features become pretty much unused and unknown.
---------
There should be a better way of showing people what features are available. If KActions supported comments about what what they do, then maybe we could do this by showing a comprehensive list of KActions in an application and the associated comments. If people don't want to even take the time to look at that list, than its their productivity loss.
If KPersonalizer would offer an "streamlined" versus "fully-featured" slider, I think that would be a very good solution.
---------
I think that may be the only way to ever streamline KDE's UI. As your agreement with some of my points about Konqueror's toolbar show, there is probably some low-hanging fruit still left. IE: stuff we could remove while pissing off only a few users. Beyond that, if KDE is ever to be friendly to those who prefer a streamlined UI and still want the power of KDE, then we'll have to find some way to allow people to have that streamlined interface without pissing off existing KDE users.
> If KActions supported comments about what what they do, then maybe we
> could do this by showing a comprehensive list of KActions in an application
> and the associated comments.
They do support this. Look at the setWhatsThis() method. The trouble is that the configuration dialog isn't very good and amongst many problems it doesn't use this information.
Rich.
> I like them tucked away, because I prefer to use keyboard shortcuts.
But would a setup like that really be perfect for all users? If you use keyboard shortcuts, you obviously use the application for quite some time and know exactly what features you really use and what you don't. A beginning user first doesn't know the keyboard shortcuts by heart and second still has to decide which toolbar icons to throw away and which to keep.
> Currently, KDE is set up to please you and displease me.
Believe me, I also have changed Konqueror a lot.
> The question is to make the UI such that it displeases the fewest number of people.
In general, yes. However there are some things which have to be taken into account:
- The defaults are used more by beginners
- Buttons are more easily removed than added
- You have to show off features to get them to be used
- Only more features get you new users, I've never heard anybody recommend a program because it can't do something.
I personally don't think that Konqueror is overdoing that altough there is probably always room for improvement. However I think that a beginner would rather like to throw away icons he doesn't use than test through all available icons and add them later.
As an example let's take the "up" arrow. Konqueror is the only browser (I know) that is featuring that and I (and as it seems a lot of other people as seen in another subthread) like it a lot. If that arrow wasn't there in the defaults, I would have taken a lot more time to find out about it's usability.
And in your example of Konqueror you left out features which are used by almost all users, for examples bookmarks or the "downloading/busy" animation.
> I would argue that KDE's current UI is not set up that way.
Well, that's a claim without any proof.
> I'm fully support leaving in configurability, even if that brings some complexity. I think that when the GNOME people say that KDE is to configurable, they're full of crap.
Finally something we can agree on :-)
> What I'm against, however, is having the defaults be complex, when HCI research shows that most users like things to be simpler.
If that were true, nobody would use MS Office.
In the small constrained world of HCI-research with their preplanned and simplified actions and tasks, simple is probably the best.
However in the *real* world with *real* users doing *real* work, which is completely unplanned, random and complex, users are always asking for more and more features. For example almost everyone I've shown browser tabs to has liked them, even people who aren't computer cracks, but browser tabs are much more complex than the simple one site = one window layout we had before.
And it all depends on what you test in HCI research. I've read a pretty detailed report on one of those tests and in that case (and I assume in most studies it's similar) every subject was testing the system for about half an hour.
Sorry to bring it up again, but for half an hour straight, your Konqueror would probably be the best in HCI-testing and would beat almost any other configuration. But in the real world it would be declared completely useless within a week as soon as the user wants to use bookmarks.
But would a setup like that really be perfect for all users?
----------
It wouldn't be perfect for a lot of users. But the HCI research shows that people only use a fraction of the capabilities of a program, and that is what people like Apple make visible by default. I don't know if this is optimal, but the UIs that really get praised for their usability (NeXT, MacOS, Amiga, etc) are like this, so there must be something to it.
Believe me, I also have changed Konqueror a lot.
--------
Does anybody like the default Konqueror setup?
The defaults are used more by beginners
---------
And beginners tend to only use a fraction of the features of a program. That argues for having the default UI only present the most commonly used features, keeping other features in the main menu.
Buttons are more easily removed than added
---------
Not really. It is equally difficult to do both.
You have to show off features to get them to be used
---------
There has got to be a better way of showing off features than cluttering the interface with ones that users will most likely not use.
Only more features get you new users, I've never heard anybody recommend a program because it can't do something.
----------
This is completelly wrong. The thing that draws millions of Mac users to the Apple platform is the elegance of the UI. And nobody is suggesting that we get rid of features. I'm simply suggesting that we not put every possible feature in the toolbar and context menu. Microsoft goes to a lot of trouble every year streamlining its UI. Take a look at the Longhorn IE screenshots and compare them to Konqueror.
However I think that a beginner would rather like to throw away icons he doesn't use than test through all available icons and add them later.
-----------
Actually, HCI research suggests that users prefer to get a hang of the basics of a program first, and then be gradually exposed to more advanced features, not the other way around. If a user wants to become more skilled at using a program, t hey will seek out new features. What we should do is make it easy and quick for them to find these, rather than burdening them with all of them at once.
Well, that's a claim without any proof.
--------
Well you don't have any proof that KDE UI is set up to displease the fewest users. I've at least got HCI research that suggests users prefer simpler/more streamlined UIs, and unless you are willing to argue that KDE is a simpler UI, than I think my position is more compelling.
If that were true, nobody would use MS Office.
--------
I don't know anybody who likes the MS Office UI. I do know lots of people who raved about ClarisWorks or WordPerfect, however. People use Office because they have to, and because their job gives them MS Office workshops. Microsoft is *not* the people we want to get HCI cues from.
For example almost everyone I've shown browser tabs to has liked them, even people who aren't computer cracks, but browser tabs are much more complex than the simple one site = one window layout we had before.
----------
Its a cost/benefit analysis. Sometimes, the benefit of a feature like tabs is enough to outweigh its costs. I'd argue that the benefit of putting "Increase/Decrease Font Sizes" in the toolbar instead of the main menu is not enough to outweight its cost.
But in the real world it would be declared completely useless within a week as soon as the user wants to use bookmarks.
----------
First, I was only analyzing the default toolbar, not Konqueror as a whole. At least in 3.1.5, from which I'm typing this, there is nothing related to bookmarks in the default toolbar. Bookmarks is in the main menu, where it should be.
It wouldn't be perfect for a lot of users. But the HCI research shows that people only use a fraction of the capabilities of a program, and that is what people like Apple make visible by default.
---
You are spreading BS by omitting the fact that while 90% of the users use only 10% of all features, every single person has an own range of 10% so in the end there is still the problem what default to choose. KDE stays away from that and instead let you decide yourself (or choose a distribution which did so for you).
Your steady reference to HCI researches is nice and dandy, how about contributing actual usability studies about KDE to the KDE developers instead so they actually change something in contrary to this needless discussion.
> You are spreading BS by omitting the fact that while 90% of the users use only 10% of all features, every single person has an own range of 10% so in the end there is still the problem what default to choose. KDE stays away from that and instead let you decide yourself (or choose a distribution which did so for you)
Yes.. however, I think that this policy is wrong, since it seems like catering to *everyone* in that 10%, and thus not satisfying barely anyone *by default*. The first time I run KDE, I have to mess with quite a few settings (like toolbars) to make sure that 90% of the program is efficient.
GNOME doesn't seem to have this problem. It has other problems of course, which makes me pick KDE as the best desktop overall.
Yo man, if a project doesn't want to patronize its own contributors with the source distribution (the only thing KDE offers) then it actually need to cater everyone, otherwise it can quickly lose a considerable number of them. Binary distribution is not in the hands of KDE, and every distribution is free to choose the defaults it thinks is best for its customers, and the customer is free to choose the distribution which offers the best defaults. What wrong with that (except that many distributions don't take the chance to differentiate themselves from others, and KDE gets the blame)?
GNOME chose to patronize its own contributors and alienated not few of them on the way. Now it seems to have mainly contributors which are happy with the default, everyone else is bound to look for something else after some time (I think eg. the rise of popularity of XFce DE is quite telling here).
> - Increase/Decrease font sizes is probably not used enough to warrent being on the main toolbar.
This feature is EXTREMELY handy. A lot of people loves it,
compare it to Mozilla or Explorer, where people sit with
their eyes 1 inch from the screen just to be able to read.
I hope nobody ever takes it away.