Slashdot Mirror
Project GoneME Fixes Perceived Gnome UI Errors
An anonymous reader writes "Project GoneME is the first attempt to try moving the GNOME Desktop into a new direction. The intention is to create a community of people, who are willing and interested to help fixing issues brought up by people for a very long time and make the vision of a usable Desktop in the means of good old Unix fashion become true. In case you are interested to help, please join the project. Plenty of people have shown interest and welcome this step and the IRC channel got filled up within a short time." Update: 07/26 02:33 GMT by T : A project mailing list has been set up for anyone interested in taking part in this endeavor.
Glad to see someone improving it, but we always have to ask the question -- how much better might things be if the GNOME and KDE teams were working together instead of separately? That is, coding/philosophical differences aside. Granted, choice is good, and it's their choice what they want to work on.
Error 404 - Sig Not Found
kde, gnome, sun java desktop goneme, how many desktops will there be before one of them becomes truly useful.. or is the linux community not concerned with this?
Actually, I could care less about such wonderful things as GUI Errors for the moment. I would just love File Types to work properly. Then again... when I add a new File Association, it is kinda fun to keep adding it over and over until I get mad and go watch TV.
Instead of fighting for one signe UI, Gnome should have two modes: beginner and expert.
beginner mode would be where Gnome is currently heading. Export mode is where us, the experts would like to see Gnome go. For instance, why not have two types of file selector dialog? The current one, and if in export mode, a new one which allows people to actually type the full path if they want to? No spatial Nautilus when in expert mode.
Actually, in any of the modes, one should be able to easily configure a feature according to the needs. For instance, maybe a beginner would still like to type a full path, so somewhere (not in gconf only) there should be an option to enable it.
Out of the box, Gnome should be made for the common user. But we should have options for the power users.
Remember the year 2000? They promised us flying cars. They delivered the PT Cruiser...
spatial nautilus. of course you can argue it both ways but IMHO and a lot of other people's, it was a step backwards.
Please change the name first. It reminds me of Windows ME.... I always liked GNOME looks'. It always striked me as the sleekest guy around for (GNU)Linux, but it always suffered from serious technical usability issues, especially when compared with KDE. Nowadays I use Konqueror as my file manager, inside good old Windowmaker. I'd love to see (a fork of?) GNOME reach a level that brings it up to date with KDE in usability issues. So thumbs up to your project. ;)
Where is that guy who'd die defending what I had to say when I need him?
This is that oGalaxy guy, right?
He's been complaining about GNOME post 1.4 for a long while, mostly on OSNews. I have no idea if the fork will succeed, but at least he's putting his money (time, code, effort) where his mouth has been.
I prefer gnome over the other choices (too many to list) because for the most part, stuff just works. However, I for one wouldn't mind seeing the ability to put different desktop pics on my seperate workspaces. Maybe this functionality is available now...If so, it's not easy to find.
The order of the buttons. I think the GNOME guys were correct in 'mimicing' the Mac button layout. I think their quest to change that portion back is a mistake.
Otoh, yes, GNOME is bloated and getting rid of the registry concept is a good one. Spatial Nautilus sux as well. Yuk.
Of course, most of the people in the IRC channel are core GNOME hackers who think this is really quite funny.
After reading this guy's site, he basically seems to want a cluttered interface. Lots of options, lots of what he's used to. GNOME is about simplicity and clean-ness, as well as trying out new UI paradigms. Spacial browsing is much better after you get used to it. But he wants it to be like Windows. GNOME is not a Windows clone.
Maybe he should try KDE instead? That does everything he wants, and has tons of configurable options. I think you can modify the Earth's rotation speed in the KDE Control Center.
That said, I'm sticking to GNOME. It's very simple and clean, and doesn't get in my way. I really love GNOME 2.6 (actually I'm an XFCE user but decided to try it out today... it's niiiice).
My other car is first.
I just finished reading the rant/mission statement on the project's home page. This looks like some guy is unhappy because GNOME doesn't quite fit his vision of what a Real Man's Unix Desktop should be, and he's ready to mobilize the entire FOSS community to 'fix' things. He seems to take some of the UI choices in GNOME really personally, too.
I'm willing to give this effort a year just to see whether the rhetoric is backed by any ability.
Reverting the button order just because inferior systems do it differently is a very bad idea.
In GNOME 2.8 spatial nautilus will be the default, but there will be a visible nautilus preference to turn it off.
In GNOME 2.6, the option still exists in gconf, but not in the UI.
So, stop whining!
Gustavo J.A.M. Carneiro
Well, they are percieved errors. The GNOME developers had a good reason for puting the primary button on the right, and don't see it as an error, thus it is perceived by Ali Akcaagac.
Yet another signature that refers to itself. The irony and humor is dead.
Well, first of all, Gnome is unusable junk. It's so slow (with "Nautilus") on my Linux machine that it's not even worth trying to use it. KDE is no better, so I continue to use fvwm 1.0 for the 11th consecutive year. Fast, stable, makes sense to my parents.
I'll probably get modded down for suggesting it, but the model for a usable desktop should be Mac OS X. Ignore Windows, KDE, and the current Gnome/Nautilus. OS X makes them all look shabby and thoughtlessly designed.
In some respects, the question of a usable desktop is pointless when someone un-technical, like my mom for example, can sit down at a Macintosh and figure out how to do everything she wants to do without reading any documentation--digital photos, movies, music, email. The desktop may be great, but the OS and its associated user-space programs *must* achieve this sort of ease-of-use if they're ever to be taken seriously by Joe Desktop.
The Gconf/Windows registry comparison is wrong. The only thing is that it contains configuration data stored in one frontend. This interview with Havoc Pennington might clear up some of the misunderstandings.
I'd recommend everyone who wants to be a part of the UI debate to read the Gnome HIG before talking - that too contains information about both how and _why_ Gnome looks and acts like it does.
I saw someone suggesting an expert mode. It has been tried, and it doesn't work. But why should we have it? The only thing it leads to is more confusion. And, there are tools in Gnome that are very powerful, yet very simplistic. Look at it this way: Most often, it's not the tool, it's the user. Having more features won't make the user more powerful. It will make the average user less powerful and confused, whereas the power user will have no problem using the simple interface. I consider myself a power user, and I've been using Gnome since 2.0. In every part of my life, as a programmer, student, musician, whatever - I prefer simplicity to advancedness. Because something simple created to perfection will always be better than something advanced. This is what Gnome gives me now - Simplicity and concistency.
This new project surprises me a little bit. It's not because it's a good thing, but because I'm amazed that this man actually has the opportunity to gain support anywhere. I always try to be objective and understandable, but in this case it's not possible: Ali, or oGALAXYo, tends to troll around on osnews, and formerly the gnome.org mailing lists, accusing people, and generally being angry, and when people tell him to stop he replies with yet more accusations of how people attack him. He's kinda like Dave on Paradise Hotel (Yeah, I've seen it a couple of times).
I have absolutely no faith in that Project GoneME will do anything successful for the Desktop users. Especially when led by a man who in one post love a part of gnome, then two days later hate it - or suddenly hates Gnome as a whole and loves KDE. Then, all of a sudden, KDE is the wrong part. I'd love to see a roadmap for this project. And I'd love to see it change every day.
First of all, it complains massively about simple things as button orders, things that users don't notice on any other plan than an intuitive one - and he says things about f.i. esound (yes, it needs to be replaced) that are just cluttered with ignorance - a sound daemon has its use, ask any distributor.
Oh, and Gnome has a bugzilla. That's the place to tell anyone if you've found a bug or feature missing.
To end this post, I'd just like to say that I'm not a Gnome official in any way. I do support and participate in the community, but many people seem to think that everyone talking about Gnome positively belong to the Gnome set of developers, and often end up talking negatively about Gnome because of things that _are not part of Gnome at all_.
How anyone figures open source is a curse due forking is only a result of failure to realise the mainstream proprietary systems have a ten year head start over open source.
And even when that gap shrinks due ten years becomming a smaller and smaller percentage over time, there is still the matter of proprietary taking from open source such ideas that it then focuses on to polish for sales.... where open source is a much larger force that does NOT deny possibilities...
About forking..... well guess what.... the good things that various forks expose can then later be reintegrated to come up with something even better than what proprietary would have been able to on its own..
Forking is just one part of a bigger picture... the other part is re-integration of good things...
Surely if these folks are wanting to evolve gnome into something else. A more appropriate and funkier sounding name would be "GeGnome" (pron. genome)
...
Nick
Electronic Music Made Using Linux http://soundcloud.com/polyp
Great! Start coding! Stop talking about how other people should innovate and take that responsibility upon yourself. Stop talking how people are apathetic and start doing something other than sitting around complaining about how people sit around and complain.
You seem to like making music, (your homepage, I assume it's yours?) so do that for an open-source project, or use that creativity that allows you to create music to help in an OSS project somewhere.
I touch computers in naughty places
Less rantish, and I agree with everything he says here.
t ml
m l
http://www.whiprush.org/2004/07/ten_gnome_nitpi.h
Oh, he also talks about GoneME. He has a very low opinion of it.
http://www.whiprush.org/2004/07/its_not_a_joke.ht
steveha
lf(1): it's like ls(1) but sorts filenames by extension, tersely
Here's what I posted a while back about this in my livejournal:
He's actually not proposing a fork per se, more like a place to collect patches to the mainline gnome that are unlikely to be accepted into mainline gnome anytime soon.
-jim
I think what all these UI efforts and all this discussion in the Linux community show is that people realize that with OSS, they can actually make a difference. And it also shows that there are many different preferences and needs when it comes to UI.
The biggest problem is the language people use to talk about these sorts of projects. Talking about "GoneME fixing perceived Gnome UI errors" is a good start. But the GoneME developers themselves should be aware that they are just developing something different for a different community, and that they aren't necessarily "fixing UI errors". I mean, the Gnome 2.6 developers aren't stupid, and they didn't set out to create a system with "UI errors" (personally, I think spatial Nautilus is a slight improvement).
With Windows or Macintosh, you get whatever Microsoft or Apple tell you is best: you can buy it or you can leave it. Complaining about usability problems with those systems is useless--the companies aren't going to listen anyway.
A house divided.
Good point, but there has been so much smoke and brimstone over his "issues" that an actual, measurable metric to see how many don't like the situation could be helpful. I can't see how a button order change could take more than a week to get over, but something must be upsetting them based on the number of ex-GNOMErs I see using KDE.
If a large number of people start using these patches, then perhaps the RedHat/Sun/Novell corporate types leading the GNOME project these days may rethink top-down "shut up, we know whats good for you" decision making. If GoneME vanishes without leaving a trace, we'll learn something about how much smoke a few arsonists can create out of trivialities.
The sooner people realize that there is no single "best" user interface and that all UIs still have lots of problems, the better for everybody. Furthermore, anything that you change about a UI is going to make some people unhappy. The good thing with Linux, X11, and its choice of UIs is that UIs really are in competition.
OK, sounds good so far. No bloat. That's why I want to get Windoze completely off my home network. But then I read on ...
WHOA, NELLY! You can't talk about simple and un-bloated software, then praise emacs as a "quality piece of software," and expect to be taken seriously.
-paul
Pistol caliber is like religion: everyone has their favourite, and theirs is the only right choice.
I prefer a simple window manager like blackbox but still
use many of the KDE and Gnome parts.
It is for example possible to start a gnome-panel within
blackbox by typing "gnome-panel" into a terminal and then
get rid of it, if no more needed. For KDE applications, one
often gets annoying messages like
" QPixmap: Cannot create a QPixmap when no GUI is being used"
and sometimes the application does not work.
I would like to see more elements of different windows managers
to coexist peacefully with others without actually have to run
a specific window manager.
I wrote a review for a now defunct publication way back when GNOME 1.0 was first released. Comparing it to KDE at the time, my review said that it would have been a toss-up if GNOME 1.0 hadn't been so unstable. Anyone who remembers GNOME 1.0 will remember just what a crash-happy bugger it was.
I liked it a lot at the time, however, and I faithfully stuck with it (over KDE) for several months.
If GNOME had stayed on essentially the same track, adding only polish, features, unity and stability, I'd still be using GNOME today.
Instead, each new release of GNOME has taken away or changed more of the things I used/liked about it (read any Slashdot story, including this one, for a users' lists of grievances) and sometime during KDE 2.x, I went back to KDE. I've continued to track GNOME releases (I've got a fresh Fedora Core 2 install right now, so I've had a chance to test the most recent distributed GNOME desktops) but GNOME continues to travel farther and farther away from where I want my desktop to be.
Meanwhile, KDE has continued to steadily improve and with each new KDE release, I find myself happier and happier with my desktop.
It's a shame, but at least for some audiences (myself being a part of them), the height of GNOME's usability and coolness was probably the crash-happy GNOME 1.0. Instead of fixing the stability and polish problems and making it a nice desktop, the developers have gradually turned it into a less and less usable environment, an environment that I always feel is talking down to me while it tries to keep me in a kind of straitjacket.
STOP . AMERICA . NOW
Please, someone, ask Sid Meiers (of Civilization) to help. He knew more about effective User Interfaces 15 years ago than anyone else seems to know today.
Even back in the days of Railroad Tycoon, he was able to present understandable and usable information at the single pixle level.
User Interface does not, and should not, mean "looks like MS Windows".
If Sid Meiers isn't willing to help (lead) the effort, then study his best work before starting.
The point is that KDE and the early GNOME both offered more configurability without having to delve into things that resemble the Windows registry.
Here is the difference in our philosophies:
Current GNOME advocates:
- Configurability means learning curve
- Learning curve = bad
- Remove configurability, users be damned
This simply refuses to serve those who are not in your majority (and I should note that I don't at all buy that this homogenous "majority" of users exists; to be confused by too many options is one thing, but to suggest that all users therefore want the *same* desktop is a huge logical disconnect).
I simply believe that a better philosophy is:
- Configurability means learning curve
- Design intelligently to minimize learning curve
- While maintaining configurability
- Thereby *potentially* serving *all* users
What GNOME advocates of your ilk are saying is "if you don't like it, don't use it, even though we once provided what you like and it would be simple and unobtrusive to add it back." Now someone else tries to add it back, and GNOME advocates are freaking out.
With that attitude, the GNOME community shouldn't complain when all of the people you've told to get lost (including app developers and the sorts of Linux users who go to *help out* at installfests) abandon you in droves. It is, after all, what you wanted. We're the users you didn't want to serve.
STOP . AMERICA . NOW
Xfce4 is built with GTK2, so it will display pretty antialiased fonts. It already has an Enlightenment-ish desktop root menu and pagers, and it can be easily configured to behave more like E as well. For example, you can turn off the taskbar and replace it with an iconbox. In my experience, Xfce4 is quite stable, and it doesn't seem to hog system resources, either.
Gnome has been going downhill ever since the switch from Sawfish to Metacity. I know that sawfish has certain 'issues', but this is hardly an excuse to switch to an alternative that is missing most of the features.
Somewhere the Gnome people got the idea that usability and configurability was a negative and their best bet was to make an unconfigurable unusable interface.
Pathetic.
There are degrees and types of expertise - it's not a linear order, let alone merely binary. My favorite "recent" Windows innovation was the "don't show this widget again" checkbox. A more flexible and open system like a Linux desktop should go all the way with that kind of flexibility.
The base widget class should include properties to represent a default value, whether the widget appears at all, and possibly a "shown default and disabled" state. The base widget container class should include a widget for managing the default values and display states of contained widgets. Then the desktop containing all those containers and widgets could have preset collections of widget states, identified as a range of expertises, and a collection of user-configurable setting collections.
It would be easy to set some apps to more expert states than others. That could be done remotely, or at login, by an administrator. Suites of apps could have "expertise overlays" which set expertise for more options in the GUI when performing operations with different sets of apps. And a learning feature could offer to hide infrequently used widgets in their new default state.
--
make install -not war
*flame mode* ...
Oh, I don't like it, and I don't like it, oh, and it is broken because of Spatial mode which I can't get to
*/flame mode*
Ok, first of all, about fork - I don't get a news. This guy gets too much attention, it is not worth that for even himself. If he will get anything done, then we can welcome him as proven his point. Until then, he is simply... a flamer.
BUT let's look at the problem from other side - fact one, there are many (however, we can't count how much percent of GNOME user base) people who doesn't like the way GNOME drives away from childishly old UNIX style of thinking (in GUI case, not in overall) and thinks that all this HIG thinky is stupid and so on and so on. fact two, many people simply dislike GNOME because of serious companies backing it - and guess what, again it is partly of HIG and simpliness/coolness GNOME provides. It's all against everything geeky, in their opinion.
So there is very practical solution - write a Control Center-like superb GNOME tweaking program for expert mode!
Or there is second, emotional solution - prove your point maybe with providing details and all info for another Usability Guide. Prove your point that buttons should be in that order you have used to use, not how current HIG suggests. HIG doesn't have to be perfect, so if you have something really to add, then do it. Don't rant.
p.s. While I wrote this post I read that someone compared Windows Registry with GConf. Sights, if they have EVER used it, then they won't be talkin bullshit. GConf rocks, I would really love that many programms of GNOME would use it. It is easy to hack, easy to use, easy to change from ssh session for client, easy to make lot of kickstart options for bunch of users. It's all very simple and useful XML conf structure, nothing of big fat one file Windows registry.
p.s.s. rembember, there are ranters and flamers in all kind of camps - GNOME, KDE, Linux, BSD, Windows, Apple, whatever. I don't hate those people, however, I hate the whole process. It's all useless.
user@ubuntubox:~$ stfu This server is going down for shutdown NOW!
When I first read the /. post, I was excited, because this is exactly what I wanted to do with GNOME as well. But after reading the introduction, I am a bit taken back by some of the phrases the author used, such as:
It's totally regardless for them what the opinion of users are, what only matters is that they must be right because they say so.
and:
I on the otherhand think that some decisions have upset quite a lot of people including me and there was no possibility to bring these problems up on the GNOME Mailinglists or the IRC channel without getting yourself trapped into ugly discussions, slandering, defaming, mobbing or even stalking.
and this:
It would be nice if they could do their own little thing in their own world without convincing everyone else that they must change their stuff the way they like because they said so.
While I agree with the project goal in general, the use of such spiteful language may drive some developers away, especially if there are some GNOME developers who want to participate in both projects. Even for me, now I am afraid that I would be signing up for a war against the GNOME project.
I know the feeling you have, being ignored and even mistreated, but the introduction of your project home page is no place to amplify these complaints.
Be positive, I believe that will win you more community support.
(referring to spatial Nautilus:) "it was a step backwards."
I disagree. The navigational Nautilus found in versions 1.4 - 2.4 looked too much like a web browser. It was confusing for both Windows users, and KDE users. When a program has a navigational toolbar along with a location field, I for one would not think it would be unreasonable to assume that it has web-browsing capabilities.
I think that there's a few issues that means that free desktops need to play "catch up" with the likes of Windows.
When a free software project starts, *GENERALLY* (not all the time) the coders are writing the code because they want/need it. They aren't coding with users in mind, they're coding something that they want and think might be useful. So the project is designed for a skilled computer user, and if usability comes after that as a result of enough requests, it is already "playing second fiddle". The reason that a certain usability feature doesn't get into the code might (but of course not always) be simply because the coder uses the desktop system, and considers the addition to over-simplify the system to the point of almost being patronising (There are many examples where Windows can be considered extremely patronising to a "power user").
Speaking of being patronising, there is also a notable point in regards to the attitude of many geeks/hackers. As the "Portrait of J. Random Hacker" says in the "Weaknesses of the Hacker Personality" section:
That, and the brief mention of "Stupid People" in the section entitled "Things Hackers Detest and Avoid" is also part of the problem. Hackers/coders tend to react very badly to timewasting tasks and stupidity, so when an inexperienced user has a problem with a current system, they tend to receive ridicule and/or abuse, rather than their concerns being taken on board. This doesn't happen in every case of course, but the most common answer to a technical question is "RTFM". It's ultimately hard to really take what inexperienced users need on board when you just consider them to be stupid for not being able to use your current system.
Another thing is really the power of the (normally Bash) shell. A lot of *nix users are people who grew up on the system before GUIs really became popular, and they have got so used to a command line system that they often shun the very idea of a GUI system. When you're so comfortable with a shell window where you can do just about anything you need to, there's less of a focus on usability of a desktop system. Provided you have a basic file browser, which is usable and functional, there's a danger of not fully developing the file browser, on the strength of the fact that you can get to where you want to go much more quickly with cd /home/blah or similar at the shell. With Windows, the command line is so utterly piss-poor by comparison (yes you can get 3rd party Unix command line apps, but on it's own, it sucks), you're basically forced to use GUI systems for just about everything.
There's also a bit of a Catch 22 situation about it. Unless you get more inexperienced users on the system, you won't get more design suggestions from the usability viewpoint. But if you don't make the system more usable, you won't get more inexperienced users.
So what to do if you don't have your own basic user focus groups like Microsoft? Well, you use some of the resarch that they have done. While UI designers have been accused many times of making desktop environments too much like Windows, at the end of the day, that is what people are used to. If you want to move a user from Windows to *nix, they will have a much better experience if they are sitting infront of a system which is similar enough to their previous system that they can find their way around with little assistance. I know that many people try to set themselves apart from Windows users (although there is a large degree of elitism about that) but at the end of the day, Microsoft have been de
Yep, although their intentions are good the gnomers obviously never read User Interface Design. Too bad...
In my opinion, (which might be different from yours, which I respect) gnome is based on a too old user interface paradigm. Because KDE doesn't have rigid UI requirements like gnome, a couple of apps have made use of a much more modern UI approach. K3Bs interface for instance is excellent.
Making gnome applications have a more modern look and feel will be a monstruous undertaking. It's great so many want to participate in this project, which could be the start of something really cool.
Please login to access my lawn
KDE is close with www.kde-look.org but I'd really like to see a DE pull it all together and create a dynamic user/developer environment.
Comments.
Ratings (good for both artist/developers *and* users)
Pictures! (eyecandish interface and background picutres! you want to attract artists and excite users!)
Oh, and no patronizing, but it sound like you got that part already!
Quack, quack.
I'm definitely in a minority here, but my biggest complaint about Gnome is the way the source code is distributed.
I've developed a perverse habit of wanting to compile large portions of my system from scratch. Gnome is a nigh-incomprehensible mass of interdependencies and it's a mess trying to figure out what the minimum set of packages is that I need for a particular package.
There IS a mistaken impression that KDE is simpler in this regard because it has "fewer" libraries, but I don't think that's true - it's just that most of the necessary libraries are collected into a much smaller set of source trees. The Gnome equivalent of QT (a single source download) is "Glib and GTK [and Pango and ATK?]". Gnome requires "Orbit" and "Gnomelibs" and "Gnomeui" and "libidl" and "gnomeprint" and "gnomeprintui" and "bonobo" and "bonoboui" and "gconf" and "gconf-editor" and "gtkhtml" and "gnomecanvas" and....probably a dozen others that I've forgotten - and if you want to compile them up "by hand" (which I often do, glutton-for-punishment that I am) you waste half of your time trying to figure out which order you need to compile them in because the interdependencies aren't obvious.
It appears that most or all of the discussion of Gnome improvements has to do with user-interface issues, though, so I don't think anyone on the Gnome side feels this is an issue.
As far as I can tell, KDE actually DOES have equivalent individual libraries to all of these...but all or nearly all of them are part of the combined "kdelibs" source package. I think this kind of coordination is why KDE is often perceived to be more cleanly "integrated" than Gnome (whether it really is or not).
I wouldn't care except that some of the individual Gnome applications really do seem to be really nice. There doesn't appear to be anything remotely approaching GnomeMeeting for KDE, for example...
Hacker Public Radio is our Friend
"[Regarding the 'cloud,'] ownership was what made America different than Russia." -- Woz
XFCE. Seriously, if he's all anti-bloat and whatnot and is all up in arms about GCONF and lauguages other than C, then he should go use XFCE.
----- Question authority, but not ours. Hate the man, but we're not him.
Frankly, it drives me nuts at times, the way that, say, Windows 95 did.
I know what you mean, but I don't think that it's just a problem with Gnome. Linux is now overrun with pretty-looking facilities that only help marginally with our ability to do useful work, and in some cases they actually decrease our overall ability by making the system more obscure.
Linux and the BSDs are primarily tools for power users, because that's what their remote ancestor and inspiration was, namely Unix. Anything that dumbs down these extremely powerful tools just so that they can appeal to Windows users or to granny is completely wrong. Any dumbing-down interface needs to be entirely additive and optional, and not given pride of place as if it were a leading-edge goal.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
The user is handed a really sweet gun, but the clip is half-empty, and the gun is jammed.
KDE: User shouts "hand me a gun!"
13,000 different guns fall from the sky onto the user's head, crushing him to death instantly.
Use xcdroast for a month and then use k3b for more than 5 minutes. Then say to yourself, which would I rather use? I'm not going into the KDE vs. Gnome debate but K3B rocks.
Everyone seems to be waltzing around the dead moose in the middle of the ballroom, so I'll perform a faux pas and point directly at it: who do we imitate when Windows is no longer a monopoly?
Imagine that Microsoft loses its market dominance (it's easy if you try) and we end up with a consumer desktop OS market of 45% Windows and 45% OSX. Imagine a bit further and envision a world with three or four competing desktops. Who do we imitate then?
If your premise is that new users will not switch to GNOME/KDE/Whatever without "substantial re-learning", then you have to imitate something in order to succeed.
Wouldn't it be much better in the long run to innovate first? I would much rather have an innovative, fresh and original desktop then another milktoast clone of whatever the computer illiterati use.
Don't blame me, I didn't vote for either of them!