I am writing an article for my blog but am too tired to continue at this point. So I thought it would be a good opportunity to ask for some feedback. I hope to get more direction for what I want to include in this article. Please share your opinion on this matter, comment on the article etc... Tnx a lot, BartOtten

======================================

Do you ever ask yourself why KDE can't acquire a marketshare of 15%? I do. I believe Linux/KDE has superiour technology, genius programmers and enough people willing to make it a success. But somehow all those technologies, brilliant scripting and thousands of people on the forums just don't do it. Why?! At my blog I will talk about some issues I think are relevant to discuss in this matter.

Why KDE fails to acquire a big marketshare part #1: We can do anything...let's do anything!

It may sound a bit harsh when I say "No, let's not" but I will explain. The world population is currently estimated to be 6.97 billion by the United States Census Bureau. Assuming everyone has an opinion on their own that would be 6.97 billion opinions. That's a lot. And the most problematic part: they differ. And we, KDE, are scared to overrule an opinion or what a user wants. Let alone what a developer wants.

Because of that I have KsCD besides Amarok, Gwenview besides Digikam, Webkit, KHTML (not to forget Gecko) and about 40 checkboxes in every program I can start. Isn't that a wonderful thing? You choose what you like most, Penny chooses what she likes most and your neighbor can choose something what he likes most. No restriuctions, it's all about freedom of choice.

Well, if you think so too or you just want a good laugh watch this : http://www.ted.com/talks/barry_schwartz ... hoice.html. Barry Schwartz talks about the paradox of choice. Conclusion: More choice does not make people happier en perse.

The thing we call freedom of choice is probably the reason people think KDE is hard to use. I want to play a CD, should I use Digikam**, KsCD or is there even more? Is Gwenview a lightweight viewer and if so what should I use Digikam for? The answer for people using KDE is easy: they are different and can do things the other can't. People not reading KDE-blogs, forums and websites don't know and probably will stop using KDE before they discover. So point A: Duplication of functionality in KDE makes it hard(er) to use.


It could be solved by installing one default application for every part of the spectrum and leaving the rest 'optional'. But then we will still run into point B: If we have 3 applications almost identical or duplicating parts of each other, the amount of developers for each program is 1/3 of what it could be! *

[add possible solution?]
[add conclusion]

-- *so far I'm done. Please think, re-think, comment etc --
-- **at this moment Amarok seems unable to do so cause of a bug. That kind of problems will be on another blog but could be related to point B. --

Firstly, there are always problems about gaining too much marketshare; people expect you to be good at everything and no project can do that.
Secondly, Michael Porter argues that you can go for volume, quality or niche; IMHO KDE has generally gone for quality - even if it hasn't always achieved it, that has been the aspiration of many, if not all, KDE developers.
Thirdly, KDE's future isn't dependent on the number of users but on the proportion of contributing users to developers; if there are too few contributing users, the developers end up not getting the feedback that enables them to develop effectively; if there are too few developers compared with contributing users, users get frustrated at the slow pace of progress.

Marketshare is never a good criterion for success; being able continuously to improve and adapt, however big you are, is.

More people using it >> more developers knowing of the existence >> more developers.
More people using it >> more interesting for companies to support >> more developers, money and users

If the percentage of contributing users stays the same, more users means more contributions and more contributions should mean less bugs

Less bugs is more quality, more quality makes it better to use, better usability makes people happy, more people using it...and that closes the circle.

However, the blog is not about marketing (yet) and at this point I am too busy contributing to finish the blog. But thanks for the comment

Choice *does* matter, at least for experienced users. New users will be lost no matter how much options and alternatives they have. "Technical users" tend to ommit one, important, fact: using any GUI needs training. Instead of limiting those who know what they want, improve documentation and expose help information (things like tooltips, kelp links), add offline (interactive) introductory course/tutorial. Also, keep in mind that experienced users value usability, speed, configuration possibilities, since they know exactly what they need. New users value eyecandy, themes, toys and - if they are new UNIX converts - possibility of showing the world how cool are they
I think KDE lacks one important feature for newbies - "revert to defaults". Many users experiment with their settings and sometimes it leads to forum topics like "My panel is gone! halp!" - small wizard for restoring some settings to defaults could ease learning and encourage to experiments.
Assuming: keep the freedom of choice for experienced users, provide informations for newbies.

I can give you one, single reason why KDE4 is unpopular: sluggishness.

There are good studies out there that indicate that even delays of a ~100 milliseconds can impact a person's perception of performance, and make them less likely to reuse the website or application that shows that delay. Little delays add up.

Meanwhile, KDE4 takes 15-20 seconds to start (with a gig of RAM and a 2+ GHz processor). The kickoff menu takes about a second to pop up, and several seconds the first time it's used in a session. Windows resize in jerks. Dolphin and other KDE applications take several seconds to render their windows. Desktop effects can sort of compensate for this (by covering up the delays with zooms and fades and whatnot), but only on computers with well supported graphics cards.

I'm not entirely sure what the cause of all this is, but I've a hunch it's mostly Plasma and its eyecandy, because everything works fine with KDE apps running in a standalone Fluxbox or Openbox session.

This is why I floated the idea of giving users a built-in alternative to the Plasma desktop. "Power users" can easily set up an Openbox/KDE desktop; but novices can't, and may not have the time to learn the ins and outs of Linux/BSD desktop configuration. And currently they're left out in the cold, with a desktop that's really, really pretty... And doesn't work very well on old computers.

Since the current philosophy in desktop design seems to consist mainly of "OMG eyecandy!" I think I'll sum things up thus:

For end users, functionality > performance >> eyecandy.

Priority 1 is the computer being useful for whatever the end user is doing with it. Priority 2 is the computer doing it fast... And way way down the line is the computer having a snazzy desktop. Eyecandy is nice, but if there's a choice between eyecandy and performance, or especially eyecandy and functionality, then eyecandy has to lose out.

Things get a bit better after disabling strigi and nepomuk.

Not by much they don't.

I give three reasons

1) Someone has a power to sell own software in every PC for people who generates the market share what matters and not how great software KDE developes, can we penetrate that power what something has on preinstallation systems.

2) People love when they do not have much of choices but they need to make only choice between two. When people can make choice do they take A or B, they take what is safe for them. The safe for them is from multiple different reasons, it can be familiar, it can be what others are using and they already know everything works with it what they have. It is about paradox of choices.

3) People do not care about others. They do not care about philosophies or matters what do not matter them. They want to live in own bubble where they only know things what makes them happy and safe in their community. People do not want to be afraid of rising a opinions of their owns and then later being pushed down by others, when it is by "dont anyone thing childrends" or "you are bikeshacker". People dont really want to think are things good with everyone else, only thing matters is that they are OK.

What we should do? There is very simple thing what would be game changer.

We need to get to situation where every PC out there what is sold, is sold empty. Without any software system in it. Every harddrive needs to be empty. And software system is not allowed to be bundled with the PC unless software system manufacturer is same as hardware system manufaturer. And this means, every PC out there should be empty, while Apple can continue doing what they do as a) they do not manufacture PC's and b) they do not license OS X to any one else.

The PC markets are not free, they are not in balance. As long as single corporation can force every person out there to buy their product as long as people wants prebuild PC.

KDE could make superior GUI, Open Source community could make anyway superior software. But the change will not come as long as single runner is given a 390 meter ahead start in a 400 meter running competition, so everyone else need to run 400 meters but this single one just 10 meters. Does anyone else runner have even a change to win this special one? No. Not unless this single runner actually wants to allow someone else to win.

It's true that the market is not very competitive at the moment, but I'm afraid I still have to disagree. Until Linux/KDE is on par with Windows 7 in the speed department, it will lose. Right now, it is not competitive at all in terms of speed.

Hmmm ... to get my KDE setup on a par with Windows, I would have to slow it down a lot. Is KDE too fast for new users?

With Vista, I could see that. Not with 7, and definitely not with XP (XP's growing obsolescence notwithstanding). Not unless you're talking about a heavily infected Windows system.

Gullible Jones wrote:Priority 1 is the computer being useful for whatever the end user is doing with it. Priority 2 is the computer doing it fast... And way way down the line is the computer having a snazzy desktop. Eyecandy is nice, but if there's a choice between eyecandy and performance, or especially eyecandy and functionality, then eyecandy has to lose out.

What is eyecandy for what there is no need by the enduser? Can you give a example?

Fri13 wrote:What is eyecandy for what there is no need by the enduser? Can you give a example?

Drop shadows. Sliding and fading windows. Most of the OpenGL effects available.

(Some stuff is useful though, e.g. Expose.)

For non-3D stuff we also have:

- "Glow" around window borders. (This does help distinguish windows, but that wouldn't be a problem if window borders were more clearly marked.)

- Lots of gradients and glassiness in the Oxygen theme.

- Plasmoids that zoom out of existence when closed, or fade in and out

- Animations galore

I don't think if all these effects Gullible Jones mentioned are crucial for everyday work. It may attract users, but also it may cause in opposite efect: users will treat KDE as a toy full of eyecandys. I totally agree with Fri13: software should be useful, fast and secure.

Gullible Jones wrote:With Vista, I could see that. Not with 7, and definitely not with XP (XP's growing obsolescence notwithstanding). Not unless you're talking about a heavily infected Windows system.

I have both Arch Linux with KDE 4.7.2 and Windows xp installed. I have no viruses or malware in Windows, but it's noticeably slower in non graphic related tasks. The same about my friends Windows 7. There are few reasons behind this. Linux kernel is simply faster and has better memory management. Firefox uses SQLite and there's always lag on Windows when the browser accesses a database. On Linux it's unnoticeable - Ext4 rocks. There are lags on Windows when there's just normal disk activity, there are lags when doing few tasks same time - doesn't happen on Linux. However, when comes to graphics Linux looses here and X is the culprit. Hopefully the situation will change with Wayland.

As for other points like "freedom is confusing" etc. it has nothing common with reality - why gnome's (don't count Ubuntu with Unity) even less popular? The choice is good. The reasons why KDE and Linux overall are still far behind are very simple - there's lack of applications that most users run. There's lack of games and some specialized apps. I'm a hardcore Linux users, but I keep Windows for the reasons I mentioned here.