Off the Beat: Bruce Byfield's Blog
Why are newer versions of free software being rejected by significant numbers of users? Three years after the KDE 4 series began, some users continue to reject it, either preferring KDE 3 or looking for alternatives. GNOME 3.0 and Ubuntu's Unity seem likely to face a similar reaction -- and they are not even in general release yet. Similarly, enough people reject the Amarok 2 releases that Clementine, a music player based on Amarok's first release series, seems to be thriving. The phenomenon is relatively new, but very real, and seems indicative of changes in free software usage that are going relatively unrecognized.
Exactly how widespread the reactions might be is nearly impossible to estimate. Because the unsatisfied are vocal, their numbers are probably smaller than they appear. Certainly, they seem less than half the total users, since many people seem more or less satisfied with newer releases. My own unscientific impression is that the unsatisfied are no more than 25-35% of all users, and may be less than that.
Whatever the numbers, there is nothing wrong with the reaction. After all, free software is all about choice. If anyone cares to keep KDE 3.5 or GNOME 2.32 alive, the licenses positively encourage them to do so.
I do not even think that such forks dilute the programming effort, since I doubt that anyone who rejects the newest releases would contribute much to them anyway. If the discontented care enough to code for the retro-forks, then I would even argue that they are benefiting free software overall by providing greater choice.
Still, I can't help remembering, it didn't use to be this way. Not long ago, people might argue savagely, but they kept with the projects that mattered to them. Something seems to have changed, and I have to wonder what it might be.
The end of the catch-up rounds
Part of the reason for the situation may be the maturity of the free desktop. Ten or twelve years ago, using KDE or GNOME meant giving something up compared to using Windows. Anyone using them did so because they cared about using free software. They were willing to put up with a few inconveniences for a while. If they were developers, they could see all sorts of gaps in functionality where they might make a difference.
However, somewhere in the last few years, the situation started to change. The major free desktops reached approximate parity, and hardware compatibility started becoming less and less of a problem. Users had what they wanted or needed, and developers had less of a chance to make a difference.
In response to this change, developers decided to try to innovate. In reading over the GNOME mailing-lists, for example, it is hard to see any particular need for a new major release, except that developers were coming to the end of what could be done with the existing release series, and wanted new challenges. By contrast, users found that existing features gave them everything they wanted -- which turned out to be not much more than a place from which graphical applications can be launched.
That, I suspect, is why some users continue to be unreconciled with current KDE development. Technically, the innovations in KDE in the fourth release series are brilliant; they include easily changed multiple icon sets, enhanced searching and improvements to virtual desktops.
But the only trouble is, a sizable chunk of users didn't care about these improvements. For one thing, such improvements take time to learn, and may require the changing of old habits.
Even more, importantly, though, the discontented users saw no need for these extra features and the slightly different ways of working and thinking that the extras imply. They were content with what they had, and many probably didn't use many of the features that they already had.
And if the new releases didn't have the features they did want -- as happened with KDE 4.0 -- well, that was just another reason to reject the changes. Never mind that the missing features eventually reappeared. By the time they did, the new innovations had already been judged, and, being content with what they had, the unsatisfied had no strong reason to take a second look.
The separation of users and developers
Another reason for the situation may be that, today, users are not necessarily developers. Increasingly, too, they outnumber developers.
Ten years ago, that wasn't the case. Eric Raymond famously described free and open source development as starting with someone scratching their own itch. That is, a developer wanted some type of functionality that didn't exist, so they wrote code that would add it, or else modified some existing . For all purposes, the developer and the user were synonymous, since non-coding users were a minority. A user revolt couldn't happen, since developers could hardly revolt against themselves.
Today, though, developers and users are no longer distinct at the desktop level. They are at the system level, because few non-coding users venture beyond the desktop, but today the majority of graphical interface users probably have no coding skills whatsoever. Nor do they see themselves as potential contributors to free software in any other role. With most of their experience in proprietary systems and their main attraction to free software being the free cost, they react to free software developers in the same way that they react to commercial vendors (and, often enough, of course, the developers are employees of commercial vendors, at least nominally).
For their part, developers still tend to see free software as centered around them. Too frequently, they refuse to listen to anyone who is not a developer, and to make decisions based on what interest them. Although exceptions exist, on the whole, free software development still has to incorporate user testing and consultation into its development cycle.
The result of these attitudes is inevitable: Users respond to changes like outraged consumers, while developers dismiss user complaints and continually fail to take them into consideration.
And perhaps on some level, the hacker culture of free software discourages developers from listening to users. If they did listen, they might feel compelled to give users what they want -- and since users already more or less have what they want, that would mean that developers would have to work on micro-improvements most of the time, which is generally going to be far less interesting than trying to write the next major innovation in desktops or user application.
The New Reality
This situation is not helped by the fact that the projects that receive the most criticism -- with the exception of Unity -- tend to be upsteam projects. A distribution has a direct incentive to listen to users because its developers want their work to be popular. But what incentive do projects at one remove from users have to listen to them? Only their idealism might encourage them, and habit may often prove stronger.
User revolts are not going to cripple free software. But they are signs of the times, and I think that more projects need to consider what user discontent says about the state of free software and the norms of development.
Meritocracy applies to users tooIt makes me smile that many of the comments here prove the point of the blog Ask yourself, if someone walked up to you in the street or at work and talked to you that way, would you feel inclined to help them?
The Meritocracy applies to the user base as well, users who have proved themselves by making constructive and thoughtful comments, who are willing to participate in the process, who can be civil and appreciate that they will not always get what they want, will get listened to. Users however do need to understand that things are sometimes not as easy as they think, things take time, we don't have infinite resources, and we do have outside lives that need taking care of (work, partners, children, other hobbies and commitments, etc). We enjoy having users, and enjoy having happy users even more, but with so many users to choose from and so little time we can't make every user happy.
I think the greater influx of users from the Windows world is part of the issue as they arrive with little understanding of how we do things in the FOSS world, and I think we fail in educating them on how it does work.
With regards to KDE and printing, it's been said a million times already but: yes it's not as good as it could be, removing the KDE3 printing features was not a decision taken lightly but we had no alternative at the time, yes its a very unsexy thing to work on and so gets almost no love, yes we're stuck waiting for Qt to make improvements Nokia is not interested in, but it works and it's not the end of the world. We're hoping OpenGovernance and OpenPrinting should see the problems solved in the next 12 months.
For the person complaining about having to change the paper size from A4 to Letter: the default values in the dialog are loaded from the printer configuration in CUPS, so change the default value for the printer itself in your printer admin and it will always be that value. You can't do that for margins unfortunately, but the app should be setting a sensible default anyway, raise a bug if they don't.
Oh, and Nepomuk. It's fair to say that there's not many compelling use cases for it so far, but it's taken a while to get the underlying foundations in place for apps to start making use of, then it takes time for apps to figure out how to actually use Nepomuk, and then they have to implement it. It's an iterative process that just takes time, and 3 years on we're now starting to see those features emerging. This is cutting edge stuff, you can't expect it to all magically appear overnight fully implemented and working perfectly.
Why KDE4 lost meMy experience as an end-user is with Kubuntu. I don't know how closely the Kubuntu implementation of KDE mirrors the KDE being discussed here, but the user who is just trying to use Linux I would submit that it doesn't matter. If they've chosen Kubuntu because they want what Ubuntu offers, but they want a KDE interface then for all practical purposes Kubuntu = KDE.
In my attempt to move from Windows to Linux I had always preferred KDE. In my experience it was much more like Windows than Gnome and I found it easy to do what I wanted. Then I upgraded Kubuntu to a new version which used KDE 4.x... I immediately found that I had a new thing on my desktop that turned out to be a "plasmoid" which showed me folders. All I can really remember about it now is that I could not work with it in any way that was familiar to me and I could not find any documentation that helped me. I also found that I could not place "shortcuts" directly on the Desktop like I wanted. Since I was fairly new to Linux and other life pressures meant that I didn't have time to spend on this, I simply gave up and continued using Windows since I was able to get done what I needed to do without thinking.
Now that I'm ready to try again to move from Windows to Linux I'll be using Ubuntu. Regardless of what the desktop experience is like I'm more confident that I'll be able to find documentation to help me on my way. Being able to do what I need to with a minimum of research effort is my paramount concern. We'll see if that proves true, but for me the lesson is not that developers shouldn't implement new "whiz-bang" things, but that users need to have a clear path laid out for them as to what has changed and how to do things in the new environment. Without this users will stay with, or go to, whatever allows them to be productive.
interestingI have to agree with this article.
I have been using Linux since the days of Red Had 7 (no not the enterprise one). I avoid code and the terminal like the plague. I use Linux because, one, I agree with the FSF. The second reason is I raised a family on minimum wage and could not afford some of the programs I required daily and I did not want pirated software. So I am a user. I used KDE almost exclusively. Loved KDE2. KDE3 was great too but now I no longer use KDE 4. Why? The big reason is that it is so slow it is choppy on a hyper threaded, 3Ghz Pentium 4. I mean really? Come on guys. Another reason is that I can't seem to be able to figure out what some of the things I see on the desktop are used for. I shut them down as desk clutter.
So now I log into Gnome on my main machine and my file backup server and LXDE on a PII that refuses to die.
MaturityThanks, Bruce, for a good article. I've never been a great lover of KDE or Gnome, but I do think it's a shame that so many users are becoming alienated from the developers, and I think it happens in other areas too. Of course the non-paid developers don't really need users at all, but it seems to me that the whole community is better off if there is a good user base with good communication.
One thing I didn't agree with was the statement that "[t]en or twelve years ago, using KDE or GNOME meant giving something up compared to using Windows." I started using Linux ten years ago, initially with KDE, then Gnome, then Blackbox, and I never felt like I'd given anything up (unless you count delays and instability). Maybe I was lucky, or just unaware of Windows' best features of the time, but I found the desktop felt much as it always had under Windows, except that now I had virtual desktops, sloppy focus and a lot more control over how things looked.
Now I'm not much of a gamer and have never even seen a copy of Photoshop, so maybe I'm unusual in not missing Windows apps, but my experience has always led me to see Windows as the 'poor relation.'
Good analysisI think your analysis is spot on, but I would stop short at saying "Too frequently, they refuse to listen to anyone who is not a developer, and to make decisions based on what interest them. Although exceptions exist, on the whole, free software development still has to incorporate user testing and consultation into its development cycle."
Why should the developers listen to anyone? Many developers are working for free. I can't imagine there is any big money behind KDE, especially with the distributors who actually take in money (e.g. Red Hat, SUSE, Ubuntu) using Gnome. Why, then, should the developers listen to what anyone has to say? It's up to the developers to decide how they want to spend their free time.
As a user who does not pay for any of his software I don't see how I have the right to feel outraged when the developers do something I don't like. I can ask nicely if they will fix bugs. I can ask rudely, too, but seeing as I don't pay the developers why should they listen to a rude person?
So, with KDE 4 as an example, if users don't like it, they can complain. The developers can listen, or not listen. But why should they do anything for users who are doing nothing for them?
I did not like KDE 4. I kept using KDE 3 for awhile, then I jumped from the big desktops altogether and started using awesome and, later, xmonad. I also keep XFCE around, and I also keep KDE on the drive because its individual programs like Konqueror are still good for some things.
People seem to think that "free software" means that they will get software developed just how they like it, by someone else, at no charge. It's ridiculous. People give away their work and then people have the nerve to get angry! If you want to control what someone does, start writing his paycheck. Otherwise, if you don't like the software he is churning out, ask him nicely to do something about it. If he doesn't listen--because he is too busy or because he just doesn't care--use something else, be it free or proprietary.
Rocks and Hard PlacesBruce,
I have found most of your commentaries to be very thoughtful. However, this time I believe you are off target. To say that, <i>"Users had what they wanted or needed, and developers had less of a chance to make a difference. In response to this change, developers decided to try to innovate."</i> is simply not true.
As a developer, I can assure you that we fight change, possibly more than users. Changing code is not only a lot of work, in a complex application, it almost always introduces bugs. Therefore, we demand a strong justification for making significant changes. However, when that justification is made, we prefer to take the lessons learned from the original development and start fresh, as opposed to building a bigger patchwork mess.
In the case of KDE (but other FOSS projects are in a similar position), the changes in version 4 were mostly driven by the Qt libraries. Having worked with Qt a bit, I can tell you that it is in a state of flux as support for mobile devices is added. And this doesn't simply mean there are new features, it also means that the handling of existing features is changing and, in some cases, going away. That is the justification that drove KDE4. You can argue implementation details, and certainly 4.0 was handled poorly. But the decision was not a whim or arrogance or wanting to show off eye candy. It was a responsible decision made by seasoned developers.
As a user of KDE and long time LUG member, I participated in a usability test for Amarok. The tester was a long time KDE team member who told us she has had great difficulty getting people to participate in usability studies. So anyone who complains that big decisions about FOSS projects are made in a vacuum, I counter with the fact that this is not by choice. The <i>volunteers</i> who develop the code are almost always open to informed, helpful comments <i>before</i> putting a lot of time into writing a lot of complicated code. But the majority of "constructive criticism" comes after so much code has been developed that it would be tremendously difficult for the <i>volunteers</i> to change it.
I posit that the gap between users and developers is not due to developers attitudes changing so much as the growth of FOSS. This has attracted many people to the FOSS neighborhood without their becoming part of the community. Prior to 1999, there was much more of a 'can do' attitude, and I don't remember seeing any complaints about developer attitudes. How many people using Linux since 2005 have joined LUGs? How many have participated in an InstallFest? How many have communicated with a developer? What contributions have they made to FOSS projects? I have detected a strong sense of entitlement from many posts (not just here) without any mention of personal responsibility or commitment to make a project better. We users of FOSS software can do better.
Later . . . Jim
KDE4 - Get back to the basicsOther comments have already mentioned this, but I'll add mine anyway. The big problem with KDE4 is they've gotten so wrapped up in eye candy like rotating cube desktop switcher (cool, but useless) that they've forgotten the basics like printing.
I don't actually print to paper much any more, but I frequently (multiple times per day) have to print to file. In KDE3 that worked great. It remembered my settings (the whole world doesn't use A4 paper, you know? and who wants 1.5 inch margins all around their page?). In KDE4, I have to spend precious minutes each print resetting all the settings to reasonable values and re-navigating to the folder where I store the files. Worse, most of the time it just hopelessly buggers up the print to the point it's useless. Instead of flowing the print inside the margin areas, it just prints wherever it wants then cuts off all the text that fell outside the margins. When printing a web page, instead of making it look like it does on the screen, it moves each frame to a different page, turning a single page into 3 or 4 or 10. They blamed all this on Qt and Nokia. I accepted that excuse and waited semi-patiently, through 4.1, 4.2, 4.3, 4.4, 4.5. Not a single printing bug has been fixed through all that. Not a single dropped feature brought back. What are the odds that they've fixed anything in 4.6? Slim to none.
Everything is slower, buggier, and with the most important features removed. I want my quick browser back. Why do I have to search my menu each time for the well hidden Dolphin, just to be able to open a file? I could go on and on, but no one that can do anything is listening. I even tried to get involved and help fix it, but I can't even get it to compile without moving everything to the latest bleeding edge unstable libraries. Sorry, but I have real work to do, and I can't trash everything to help them clean up their mess.
I've been a loyal KDE user for years. I never liked the attitude of Gnome developers, with their refusal to allow any options or customization, because they know better than we do what we need. But the KDE team are really trying their hardest to push away the last of their users.
Users are Stakeholders@Aaron Seigo
Larger open source projects are more than a Meritocracy. If everything is based on merit then the only stake holders are contributors that are usrs. Thus your only users are a small group that contribute.
Once you grow beyond that one of two things will happen. The first possibility is that you consider users who do not contribute code, documentation or debugging to be stakeholders. Lets call this the consumerisim model. Much like a local band that plays for free or an underground comic. Consumers who do not produce but consume for "free" have a synergy with the prodcuer. Which is why they produce. The reality is you can't make everyone happy. But if you know who your audince is, you can do a good job at making most of them happy.
The second possibility is that you do NOT consider end users. In which case if the decisions you make create a product they do not like. The users will comlain and then eventually leave to something else that better suits their needs, or at least does not piss them off. At which point you have a small project again that pretty much only contributors user. If KDE wants to choose the route of obscurity, their project leaders are well able to take them there.
When you say that people using KDE3 were "nice and cool", it just goes to show that second mentality. Contributors matter, and any users are incendental.
I can say me and my wife never participated in the "incivility". She desires a desktop with a transparent kicker, and wallpaper rotation. She just kept going back to KDE 3.5 till KDE 4.2 was came to be.
I try everything. So far that means I spend about 60% of my time in Fluxbox. I keep looking for something that works better, even if I have to learn more and write some scripts to leverage that power. Thus far the sementic desktop that KDE 4 aspires to be has not produced anything that works for me. Someday it may be cool if I can have a folder on my desktop that contains all files modified by my employees in the last 48 hours. So far, what has been produced has not proved useful.
As a Fluxbox user, my ability to place dock apps in a slit that rests at the side of the screen has proven more useful than most plasmoids that reside on the desktop. Fluxbox lets me group windows, only some KDE thems allow that. Plus Fluxbox will let me determine which virtual desktops apps show up on, what size and localtion they will have and how the group. All happening automaticaly. Which seems to be the idea of behind Desktop Activities, except for the fact that these "Activities" take more time to set up and work with than features Fluxbox has offered for a decade.
To give credit where credit is due, at least a properly written plasmoid can work both as a tray app and as a desktop app.
As far as good choices? It depends on where you go in the long run. Like I said, I am willing to change how I di things if it improves my workflow, desktop exeperience and adds some eye candy to boot. But you had better deliver on features that actually do something for me. So far Activities, Strigi and Nepomuk have yielded nothing of use to me. After 3 years they have produced nothing that is compelling. At some point this bloat and instability (I could buy a new car if I had a euro every time nepomuk has crased on me) had better pay off. If not you will lose more of what you consider "nonstakeholders" until KDE is marginalized.
And "probably gained" is a pretty poor metric. I want you to suceed. So far the only strenght I have seen is QT4 and the flexibility of where you can place a plasmod.
What realy amazed me, is it would not take that much work to "throw a bone" to the disgruntled users. How hard would it be to add the collapse button to the taskbar? Or to give an option to get rid of the cachew? It took 2 years of constact complaints to get the option of the old desktop back. I think many of us are bothered by the fact that it would take so little to add back in features we had before. Why not add them as no-defaults. We can get to them if we want, but highlight the new features and what you consider the "better way" of doing things?
KDE 4.X simply isn't neededI loved KDE 3.5. It was simple, easy to configure, not too hard on resources and it did pretty much what you would want any good desktop to do.
Then the KDE developers came out with KDE 4.X, which was not simple to configure, bloated and offered a bunch of 'features' that I neither wanted or needed. In my, probably biased, view, they took something that worked and broke it. I think the KDE developers suffer from the same syndrome that Detroit does.
I'm not someone that doesn't deal well with change. Change is fine, but it needs to be for the better, not just for the sake of change. I've used Linux (Slackware) since 2000 and used KDE for most of that time but I use Linux Mint LXDE on my desktop these days. It's not perfect but its fast, fairly configurable and does what a desktop should.
Still on KDE 3I'm still running KDE3 and have been actively maintaining the KDE 3 packages for Mandriva and am currently looking at Trinity. I would love to be able to use KDE 4, but I can't. The reasons are mostly quite straight forward and it would not have been difficult for the KDE devs to accommodate my wishes without breaking any of the new paradigms in KDE 4. I haven't looked at 4.6 so these comments may be out out date, but a very quick summary:
1) Performance. It still runs too slowly on my old hardware.
2) The task bar hide button is still missing (I use this a lot, auto-hide can't read my mind so is of no use)
3) Why do I have to install a third party plasmoid to hide the useless cashew?
4) Adding/changing icons on the task bar seems to be impossibly complicated.
It's not just resistance to changeWhile I'm sure the developers have a point that SOME people reject new versions purely because they're different, that seems to have become the skirt behind which they hide whenever there's ANY criticism of their releases.
The fact is, most of us who give up do so because the developers release software that no longer actually does what we need it to do. Several posters have mentioned the ridiculous mess that Kmail has become since the devs buggered up the address book. This (after trying with it since 4.1) was the final straw that made me give up with KDE completely.
An insight into just how detached the KDE devs have become from the whole point of developing software for people to use can be found at http://userbase.kde.org/KAddressBook_4.4
"KAddressBook is a work-in-progress. That doesn't mean that it's unstable but it does mean that it's not complete. The layout you see in this version is very different from the older version, and it's quite possible that the next version will be different again, as we see more features being available to us again."
Which is another way of saying,
"We had a perfectly functional addressbook, but we've scrapped it and replaced it with a new one that won't do the things you used to be able to use it for. Oh, and by the way, we're probably going to break it again some time in the future."
And the developers' response to being criticised for releasing un-functional software on the KDE forum was to blame the distros for being so thoughtless as to package up software they had released.
It's a shame, because Kmail is undoubtedly the best email client available, with functionality that doesn't exist in any other client. But with a broken address book, and worse still a broken address book that by the developers' own admission is going to get MORE broken in the future, it's totally un-usable.
This isn't resistance to change per-se, it's resistance to regressive change.
I am glad thatI am glad that we have other choices. I do not mind change but do not like change for change sake. I also do not like bloat and using resources just for the desktop. As I get older my needs get less and less and what KDE is offering is more and more. I am glad that they are adding users but for me I am looking else where and fluxbox and others are looking better and better all the time.
Devs, Users and StubbornnessThe Gnome Devs are just as dismissive and aloof from Gnome users as the KDE Devs are.
The attitude is "if you want to change something learn to code and change it yourself, otherwise shut up and be grateful for what you get for free".
The other factor is that a lot of people are stubborn and refuse to learn anything new; "you changed something in my GUI, I'm going to leave" is the oft heard cry. These people think nothing of learning to use a new PVR or a new car, but if you change something in the desktop they go nuts. They are vocal and thus appear to be a greater proportion of the user base than they really are. Becoming enslaved to their shrill demands and easily upset natures is a fools game. This might be why, at times, the devs appear aloof and uninterested in complaints and why, with some justification, they say "learn to code and change it yourself". Many Linux users are stuck in Windows mode, whereby you get what you're given and have to put up with it. Open Source is not like that.
Part of the OSS idea is that you take some responsibility for yourself and stop expecting everyone else to bottlefeed you.
For the record - KDE4 on Slackware 13.1 is perfectly stable. The first few instances of KDE4 were awful - buggy, half finished, poorly implemented. That's all changed now. But, as with all things in life, the saying "once bitten twice shy" holds true. If you tried KDE4 early in its life then you probably thought "what on earth have they done to KDE?" and ran away as fast as your legs would carry you (as did I). But, it's OK to come back now.
KDE4 devs are so enamored by their 'incredible' infrastructurethat they spend all their time extending it and playing with it but they don't actual try to use it.
In KDE 4.5 in fedora, it took me 10 minutes to get wireless networking to work!
I'm not a gnome fanboy but it took 10 seconds to figure it out in Gnome.
I keep hearing people at planetkde raving about the cool new icons finally bring the desktop into the new millinium or how cool and easy it is to post to twitter using a plasmoid or other similar things.
Do you honestly think I give a rats butt about twitter posting when it I have to pull my hair out to figure out that monstrosity that is knetworkmanager?
A perfect example of KDE mentality is the effort that went into adding 'Software Compilation' to the name and explaining why that better describes all the components, yadda yadda yadda. Seriously?!?!
As others have noticed KDE people will 'spend' 'copious' 'amounts' of 'time' (seriously, WTF with all the quotes?) explaining why the latest paradigm shift enables synergizing with social media and accelerates assimulation of new possibilites, but they won't spend 10 damn seconds saying 'yea, go to this menu and click blah to disable foo.
KDE should be renamed to PHB Seriously, you won't talk to us users like real people. It's like you are marketing people who are afraid of realizing we don't give a shit about feature whizbangyadido when I can't even open my gmail calendar in kontact or select 2UP from the printing menu.
I've filed more than a few bug reports with KDE, going back to 3.3.something. I feel like I wasted my time.
I'll say this in simple English:
Networking, printing and sound are more important than strigi or nepomuk or ktwittermoid.
Really?Really? I switched from KDE 3.5.10 on Mandriva 2008 straight to Mandriva 2010.1 KDE 4.4. I upgraded to 2010.2 recently and I am eagerly awaiting KDE 4.6 in the KDE repos.
I had no issues finding most of the KDE 3.5 features that I used to use, straight away and set it up. Spending around an hour playing around with the settings, all good. Not a single app crashed since the 2010.1 release. Dual screen setup with NVidia drivers all works fine. The same "familiar" KDE goodness is still there, and who better than Mandriva to polish it ready for mass consumption...
Can't deal with ChangeIf you make any Changes that i'm not used to, I just simply am going to jump off the end of the World.
I for one like KDE, sure it's different, but it is a better Desktop.
Thanks Mr Seigo...for confirming what many have been saying: that the KDE4 dev team, or at least its leadership, are arrogant and dismissive of KDE users. "They began to consider themselves as stakeholders. In a meritocracy, that's a pretty shaky claim to make, but it's understandable given the sort of passionate individual that was attracted to alternative operating systems up to that point...some started to feel like they were having something "taken" from them and began behaving that way." Um...just wow.
I'm going to spell out a few points for completeness sake:
- You're darned right users are stakeholders. Who are you writing software for?
- How do you think users become contributors and supporters, from absorbing heaping doses of your scorn and arrogance? Usually this is counterproductive.
- Even if many users never become significant contributors, devs are still answerable to users. Again, who are you writing software for? Obviously you do not believe this. You're emitting the same old tired and wrong complaint that KDE3 users are defective, afraid of change, and sadly unappreciative of your greatness. Good luck, you're going to need it.
KDE debian based distro choices limitedOne thing I've noticed lately is that a lot of the negative KDE reviews seem to be from those users coming from Ubuntu, or other Debian based distros.
While Debian is awesome and stable, there just doesn't seem to be a solid KDE version of it if we are to believe all these users and their problems (I doubt anyone is lying). I on the other hand have been using Kubuntu for a few years now, and I just don't have the troubles that others seem to have.
I know many suggest openSuse, Fedora, or Mandriva, but I just don't like rpm package management, and I think that others coming from a Debian Gnome based distro that decide to try out KDE will not only have to learn KDE, but will have to learn a different package management system, a sort of double-whammy. Perhaps just to many differences to get comfortable with.
User Rovolts and openSUSE RevolutionI could not agree more with you about the subject you have proposed. Although it is hard to put a hand on the dissatisfaction of Linux users over their desktop UI, I believe that this unhappiness is very real. I actually blogged about a possible explanation of this failure to understand users on my own blog. I would love to have you read it and see what you think about it.
Thank you and great post.
It's a pervasive problemIn the 3 years since KDE4 was launched, I've seen no indication that the KDE developers have ever heard the objections that I see daily.
I'm glad to say that Trinity-DE, the KDE3.5 fork, works well enough to be used, I'm using it right now. This is in comparison to KDE4 which remains so completely divorced from "just work" that I cannot stand it. Every once in a while I'll do a VM with a clean KDE4 install (Debian), and each time I delete it after several hours out of total frustration and disgust.
I'm one of those who has used KDE since version 1, and each time KDE had a major version change things just worked better. Users were happy to have it, enthusiastic even. But not KDE4. Even the most glowing report of KDE4 that I've read was a matter of "finally getting it to work like KDE3 by turning absolutely everything off."
How could any developer not realize that Kmail with KAddressBook is an integral and essential part of anyone's desktop that uses KDE? And until KDE4, it WORKED. It worked WELL.
To make a release that does not and cannot work at least as well as the previous version is a slap in the face to users. Yes, we saw that with VISTA, and it's a sad commentary to say that KDE4 and VISTA have things in common.
Sadly, the KDE developers won't even read this. I only hope it isn't too hard to get my contacts out when I finally give up and move to another mail program, which I will have to do if Trinity-DE doesn't get some active forking going on.
only half the numbersIn this article, I think you've only calculated using half the numbers. I won't speak for other software products, but let's look at Plasma ...
... large numbers of users also rejected the KDE 3.x desktop. We had a very select and specific sort of user with KDE 3 and were facing falling significantly behind the state of the art unless we made some choices. We did our best to decide what those choices would be, and then went through with those decisions even though they involved a lot of hard work and sometimes even risky experimentation.
The goal was not only to produce something that existing KDE 3 users would appreciate, but something that those who had rejected it up to that point might also consider using. As a result, we have a significant number of users who came to Plasma that never "bonded" with KDE 3. A user's choice awards survey last year had KDE's desktop offering tick 10 points upwards as a result.
The term "bonded" is key here, I think: many people bonded quite deeply with KDE 3. Which is nice and cool. They began to consider themselves as stakeholders. In a meritocracy, that's a pretty shaky claim to make, but it's understandable given the sort of passionate individual that was attracted to alternative operating systems up to that point. Between some shaky starts and missteps made by various players in our shared ecosystem, some started to feel like they were having something "taken" from them and began behaving that way.
Unfortunately, it culminated in a lot of shrill replies and sometimes just outright nastiness. The user community failed to self-moderate and, operating from positions of entitlement that are illusory, some of these individuals managed to do a hell of a lot of damage to the developers, and as a knock-on effect, the users of the software. You don't have to agree to everything, but we should be able to expect civility. Another unfortunate symptom was how many people put on rose colored glasses and claimed that KDE 3 was, indeed, perfect and had no meaningful flaws. While I accept there are probably individuals who felt that way (can find that with nearly any product, it turns out), it was not a general truism. It was good, but it wasn't as great as we could make.
At this point, we've probably gained more users than we lost and are set nicely to continue growth, both in terms of going from strength to strength in the technology and participants. That's the other half of the numbers: the gains now, versus the losses in the past.
In that light, the "revolting" users are suddenly not the center of the universe and the definition of the current trend. Yes, we could have held on greedily to the users we had while many others slipped through our fingers .. we can all ask CDE how that worked out for them (among countless others, Free and proprietary).
Hard choices? Yep. Good ones? So far it seems like it is working out. Why do we still see articles about how much gnashing of teeth there is, when we're also getting reviews like this:
? Perhaps because it's a more exciting story, or because if you complain loudly it is noticeable.
Some developers "lost" me a while backI started using KDE4 around the second alpha of the 4.0 series... Used it ever since up until a few months ago. Once the semantic-desktop and all that started becoming a dependency, I couldn't stand using KDE anymore.
I haven't found any useful Plasmoids that I would ever actually want on my desktop, and I have no need for Strigi whatever sucking up my computer to catalogue every file in my system (including my 30 thousand plus MP3 files). The few other files I have on my system, I know exactly where they are and what they are and I don't need to rate them or tag them. I also didn't use "activities" and chose to use Compiz-Fusion simply because it worked better with my graphics card and was faster and more appealing. (the background in the Expo plugin for example.... small, but appealing)
So I gave up on KDE4, and started using Gnome again. I see no turning back, but who knows how well everything will work once Gnome 3 is out?
I gave the developers of KDE4 a REALLY long chance, and they failed to really sell to me whatever it was they were doing. The entire direction of the environment seemed to be FORCED in a direction that none of the users really understood or even wanted. I don't know a single person who wants a semantic desktop or would ever have any use for it. I also don't know anyone who needs an OpenDesktop widget/plasmoid (especially since they lack a Facebook one).
Likewise, Amarok2... I used that for quite a while until a couple releases ago. I still don't know why half of the context area is even in there, or why it uses so much space. After much resistance, I am now happily using Clementine. (after a stint with the dependent-heavy Banshee that would be incredible if it didn't take so long to scan my entire music catalogue.... it took 4 days.)
All that whining being said, I appreciate that many of these developers essentially give up their lives to be doing this and making all this free software.... but they lost me on where they were headed with it.
I'm still at a loss as to why anyone would EVER need a semantic desktop.
User Revelts -- So trueYour point is well taken.
While 10 or more years ago which desktop you used, KDE or GNOME, was something of a holy war that attracted flame bat by the thousand today it's more like "Gimme something that works!"
The current version of KDE is stable enough and works well enough, using Mandriva, that I use it every day in dereference to GNOME. It''s not the old debate about KDE not being "free" enough or GNOME being "free" as defined by Richard Stallman. It's just that I prefer it to GNOME. So there!
This isn't to say there aren't problems that the developers just don't seem to get. The current version of KMail has broken mailing lists and no matter how many users pile on the devs they just don't seem to get it that these lists are central to an email program and matter to users far more than the semantic desktop does. Both are important but mail lists far more important. The problem seems to be, if I understand it correctly, that Kontact, KMail, KAddressBook and a couple of other apps were chosen as the hothouse for the coming improvements to the semantic desktop because the devs consider those apps so central. True enough but they still broke a vital part of KMail to get this far.
I have other, fairly minor, problems with KMail such as the lack of icons when you mark a message important and stuff and changing it to the mail subject line in a different colour.
Hint, icons are easier to see and the brain latches onto them better than it does to coloured text. Particularly in this day and age when coloured text in HTML email is so abused that different colour text is meaningless.
Still, I sigh and remind myself that it's better than breaking darned near everything with Vista did and having extortion used to get what ought to have been a free bug fix (Win7).
The KDE and GNOME development communities would do better to take users into account and, heaven forbid, do some form of focus grouping so that silly things like breaking mailing lists in email programs, no matter the reason, just isn't going to fly. No where, no how. Then, if they have to, work around it. The devs not the users.
KDE4 is the new WindowsThis new 'semantic desktop' guff may be all brilliant, and possibly useful to actual users. Though nobody can quite explain the benefits to users, which makes me think there aren't any. The biggest problem I see is this whole enormous dependency infrastructure they're building, to the point that running one or two KDE apps in another environment is going to be difficult or impossible. Anything depending on kdepim, such as kmail and amarok, is going to require at least 100+ megabytes of dependencies, plus a huge batch of Nepomuk and Akonadi processes. That's an awful lot of overhead for a couple of apps. It's a lot of baggage even inside KDE.
The time-honored Unix philosophy of small, specialized apps that each do one thing well, and interact seamlessly, is a brilliant concept that has proven its worth. KDE4 breaks this with no clear articulation of the benefits of doing so. I see more bloat and less flexibility. This is not good.
KubuntuThe reason why users complain is that they use Kubuntu which is broken by default and want to use a debain based distribution with its rich software portfolio and easy administration. In other words, we need a decent Debian-KDE distribution or a "RealKDE-Desktop" fork of the crippled "Kubuntu-Desktop" for Ubuntu.
changesSometimes the reason is the new versions are substandard or hardly usable (like I understand KDE 4 was initially, and like "Unity" is).
Other times it's the same as why people find it hard to switch from Windows - most people don't like drastic changes. I think it's safe to assume that in reality a minority likes to explore, learn, experiment, change. Most people like patterns and safety.
Btw, I think the new Gnome will fare better than Unity. I don't care either way as I use Fluxbox but that seems to be the direction things are headed to.
You missed one reasonOr rather, you missed a piece of a reason you gave.
I think you're right about developers continuing to develop just to have something new and interesting to do. And you're also right that many users stay with the old version because it does what they want and they don't see a need to relearn. (Windows XP owes its continued zombie existence to this same thing.) But there's another reason that goes along with that: sometimes the new versions actually reduce functionality or usability.
Your examples are excellent illustrations of this. Take KDE, for example. I would love to use a KDE distro. KDE4 is beautiful... if it would just <i>work</i> consistently. But I haven't yet found a KDE distro that is as stable as Gnome on my laptop. Every one I've tried has been pretty, and unbelievably configurable, but not reliable, and hence less usable. So I stay with Gnome. For the same reason, I'm still using Amarok 1.4. Amarok 2.x is full of all sorts of flash-bang features, but the features I actually <i>use</i> work better and more reliably in the old version. I don't want the tracks from all my "Various Artists" albums scattered around the collection list; I want them under Various Artists, by album, the way 1.4 sensibly did it.
Gnome 3.0 and Ubuntu Unity show all the signs of doing exactly this kind of thing. Take a mature interface that's reached the limits of its extensibility, and start over from scratch with a new paradigm. You can already see from the early user reports that Unity is going to be missing a lot of current Gnome functionality. Will there be cool new features? Sure there will, but there's no guarantee they will make up for what will be taken away.
I don't have a problem with advancement for the sake of advancing. I <i>do</i> have a problem with retreating for the sake of advancing. When something is made less functional just to give the developers something new and interesting to do, they lose me. It's not a revolt, just a quiet stepping off the bandwagon.
Very interestingI think that your analysis of the current situation couldn't be more accurate. Thanks for voicing the dilemma that user of free software have today.
But if you are not using the latest Linux kernel, your system is insecure.
Home routers will give room for custom firmware but still comply with FCC rules
Frank Karlitschek will continue to lead the open source ownCloud project
“Xenial Xerus” comes with a new packages format and several improvements for the enterprise.
Linux users can now download and install the Windows code editor
New initiative will address security and interoperability concerns around container technology.
Developers can use RHEL as a development platform without a subscription fee.
Windows users will soon have native access to the Bash shell.
Improvements to SMTP will provide better guarantee of confidentiality
Graphics vendor embraces new reality in Linux graphics