30 September, 2008

And how's life...

Hi web,

Haven't blogged for a while - been very (VERY) busy with work and other stuff. I've recovered from the car accident, and so has Kenny. I even bought a new car. This time not yellow (somehow that color doesn't feel right for a car anymore), but boring grey. Aaah well, at least it's corporate-compattible...

Ow, and I got a new assignment, which is interesting. I now work at a small organisation (500 employees) which distributes 2.5 billion euros to the schools & universities in the Netherlands. Yeah, each person working here distributes 50 million euro a year on average ;-)

On to KDE related stuff - a lot going on. We're discussing freedom of speech & rules of engagement - that's one heck of a topic. I always find it hard to participate in such discussions, as I generally see how both sides have a good point. And when I try to keep it simple, I have to retract or say I was sorry for what I said... Meh, I'm not good in keeping a fight going. I don't think that barfight with Sebas is ever going to happen.

Aaah, and there is the report about social innovation in KDE. Sorry guys, haven't had much time to spend on that yet, but I will. I've got a nice piece of theory ready, giving the basics - but it's still very much company-oriented, and I want to make a KDE specific piece. Thus still work to do.

Last of all, can I say KDE 4.2 is going to rock? I've been participating in some discussions here and there, most notably a dutch Mandriva forum but also at places I gave a talk about KDE. The feedback is pretty much what we've been hearing a lot: looks cool, probably a lot of potential, but I miss *pet feature X*. Why did you remove that feature etc etc. And often the "you guys are pulling a Gnome on us" is thrown in.
Hehe, they really fear the Gnome, apparently :D

But from my experience in Trunk I can tell them most issues are fixed, most features are implemented - and of course I can explain what we are TRYING to do.

To re-itterate (feedback welcome):
The reason for the missing features is not that we actively removed them, but all the work we put in the infrastructure. We heavily improved the inner workings of KDE, and now have to port, sometimes even rewrite everything. It'll take a while, but it will make KDE more maintainable, improvable and of course cool.

Then, after more questions about missing features and /me talking about 'we want to improve usability':
Don't worry, we're NOT pulling a Gnome 2.0 on you guys. Yes, we want to improve our usability, but in a Better Way (TM). If adding a feature would lower usability, we try to come up with a way of doing it SMART. We like having our cake and eating it too, and in many cases, it is possible. But when it is not, we DO implement the feature, unless it's really really obscure and almost nobody will use it. Of course, doing things this way takes a bit more time and work and experimentation. So missing features are not due to the KDE developers going the 'dumb-it-down' way, but due to a lack of time.

Meanwhile, I always mention:
Creative ideas on increasing usability while not decreasing featureset, or even increasing the featureset without a decrease in usability are MORE than welcome. Even if we're talking wild ideas here - we're willing to try things, we know that's what it takes to innovate. You've seen us do it - and we'll continue doing so.


The good thing is our users trust us, and they are often pretty big KDE fans. They do get confused by all the apparent focus on bling & clean interface in KDE, and the lack of features. But when we explain the underlying reasons for the changes, they generally understand it. And they are willing to wait a bit more. Many DO complain about the approach distributions have taken: "If you guys tell us KDE 4.2 is the release we're all waiting for, why do many distro's switch to 4.1?". I can only tell them I think the distributions are imho making a mistake - but who am I to tell eg Ubuntu, Mandriva or Suse to change their way?

What does surprise me is how big the influence is of a few missing features compared to all the cool new stuff, which seems mostly ignored. This I think is a clear proof that the few features Linux misses, compared to Windows, really hurt a lot - despite all the things we do better. Of course, part of this is familiarity as well, something Gnome does just right: incremental improvement. Unlucky for them, it gets you only so far.


Wow, that was a huge braindump, without any pictures, so I'll stop right here. Let me just finish by saying I think this community is great. The recent discussions about freedom made me realize how much we all appreciate it - something binding us together, imho. Despite (often actually fairly minor) differences in opinion.

14 comments:

  1. "but due to a lack of time."

    as you note, the approach we are taking to some things means it takes more time. so it's not just a lack of time, but the aim for quality adds to amount of time required.

    most people are willing to trade a bit of time for a bit of quality.

    "distributions are imho making a mistake"

    4.1 really is good enough for all the more conservative users. 4.0 was a questionable move, but all those distros also offered 3.5 as well.

    4.1 is a bit different: it's got enough fit and finish for daily usage for the majority of our users.

    most distros still offer 3.5 as well.

    the drop of 3.5 seems to be timed with the arrival of 4.2, so i think the distros are really on track again here.

    "all the cool new stuff, which seems mostly ignored"

    only by those who have that One Feature (or two, or whatever) that they can't mentally get past.

    "This I think is a clear proof that the few features Linux misses, compared to Windows, really hurt a lot"

    yes; though it's not really features so much as task equivalents. as long as there is a way to accomplish the same kind of task, we're usually ok. it doesn't have to be the exact same feature, though that incurs more retraining/learning.

    "something Gnome does just right: incremental improvement"

    let's be fair: we did 6 years of incremental improvements in KDE3.

    we punctuated that with a Big Jump in birthing KDE4. two important things to keep in mind: KDE3 was and is still there (so this is a good example of incremental improvement and big jumps at the same time) and that we are now back to incremental improvements (though at a really high pace).

    the issue is not incremental steps vs big leaps, imho, as both have advantages and disadvantages.

    innovation, relevance and customer retention require both to be exercised in proper amounts.

    there's some good literature on this concept out there in the management/business genre; some of the stuff that draws on Toyota's processes is really quite interesting here (they practice both incrementalism and high innovation, but in a controlled mix).

    in retrospect the thing that we can do better the next time we do a Big Leap is to ensure that those who are best served by incrementalism don't join us for the big ride quite so early on.

    still, kde has YEARS of solid, incremental improvement releases to show and the LAST people who should be forgetting to tell that story is we ourselves!

    GNOME has the opportunity with Gtk+ 3 an GNOME 3 to engage in an important Big Leap moment. if they don't, they will be stuck with incrementalism, which is fine in the short term but not enough over the long haul. it will be interesting to see how the next 2 years shake out for them in this respectd =)

    ReplyDelete
  2. Maybe a hard question, but then again, maybe you know. How much of that 2.5 billion euros is going to microsoft and now much is going to other software vendors?

    ReplyDelete
  3. I use KDE4.1++ now (trunk-packages from openSUSE) and I really like KDE4.2 a lot. Can not imagine going back to 3.5....

    KDE 4.2 is ready for prime-time

    ReplyDelete
  4. @aaron: "4.1 really is good enough for all the more conservative users" - yeah, "good enough", right.
    You keep repeating how things are not as good as they should because you strive to quality and whatnot; yet whenever someone else says about it, you'll claim that things are indeed good enough right *now* (like this good enough 4.1).
    If 4.2 is really as good as it's said to be, then it will be a first real good KDE release, and it's not bad, even if 4.1 sucks! There is really no need to fool people by "4.0 is good enough", "4.1 is good enough", you'll get enough beta testers without false advertising.

    ReplyDelete
  5. @anonymous1: 90% ;-)
    seriously, I have no idea, that's up to the schools...

    @Aaron: 'good enough' doesn't cut it for many users, as ygven notes - he has a good point...

    ReplyDelete
  6. @Yevgen: we were very clear about 4.0, so don't put words in my mouth, thank you very much.

    as for 4.1 being good enough, that's based on actual user feedback. i'm not particularly inclined to debate when i can rely on such data.

    as for things being good enough now and yet striving for more, it's really not a tricky concept. there are some things like panel hiding that have taken time; but that's not something that everyone needs nor even wants.

    "then it will be a first real good KDE release"

    yeah, all those kde3 releases were complete shit.

    *rolls eyes*

    whatever.

    i'd be more interested in hearing your wisdom about incrementalism and large leaps, how they interplay with innovation and maintaining audience. that's an actually interesting topic.

    ReplyDelete
  7. @aaron: http://aseigo.blogspot.com/2008/01/talking-bluntly.html, Meme 5 (first hit in google for aseigo and kde-4). Very clear message, indeed.
    User feedback, are you kidding? Or do you mean the selected users feedback (rhetoric question, indeed).
    And I can't really share my wisdom on buzzword and buzzword, how they interplay with buzzword and buzzword, that's a job for you, I am a mere user who does his stuff on his KDE-3.5 (I meant first good KDE-4 release, KDE of course has had many good releases before 4.0).

    By the way, about panel hiding. How about the damn old desktop? You won't explain my wife that her folder isn't there anymore because of some conceptual thingies and bingies, will you?

    You really should run for president or something. Whatever.

    ReplyDelete
  8. @muntyan: *sigh*... more disinformation, *yet again*.

    "By the way, about panel hiding."

    If you read Aaron's blog posts, you will see that it's done for 4.2, marked as done on the feature plan even.

    "You won't explain my wife that her folder isn't there anymore because of some conceptual thingies and bingies, will you?"

    http://userbase.kde.org/Plasma#I_heard_there_are_no_more_icons_on_the_desktop_in_KDE_4.1...

    http://userbase.kde.org/Plasma#What_is_the_Folder_View_applet.3F

    Those answers aren't new: I wrote those when 4.1 was released, or even earlier. You may not like folder view, but please don't say that there are no more icons on the desktop.

    ReplyDelete
  9. *yeah, all those kde3 releases were complete shit.*

    Now you can say Aaron is getting edgy :-).

    It actually looks like users expected a Win2K to WinXP style seamless transition. When KDE4.0 turned up with few front end features missing, it was seen as KDE developers biting off more than they can chew. Some users who saw the missing features as KDE migrating to Gnome / Nazi style purging of features added to the chorus.

    Given a choice between innovation with short-term disruption or status-quo with incremental changes, I'll take the innovation approach anyday. But I guess the question of wether 'KDE4 transition *could* have been handled in a smoother fashion' will never go away.

    Maybe there should have been simple tutorials shipped to aid the transition before distros made KDE4.x their default DE.

    ReplyDelete
  10. @Einar: this folderview thing is indeed a good example of how something can be simultaneously be not ready and be ready.

    So, icon position saving, it's a feature used by every person who actually uses desktop to store and access stuff on it. Is it there in 4.1? I thought Aaron said it was finally implemented, but after 4.1.

    How about this: "In KDE 4.2 you will be also able to use the Folder View as your desktop, replicating the "old style" paradigm." - that's written at the web page you posted link to.

    It's so very simple: don't cheat! KDE-4.1 isn't good enough, KDE-4.2 will be good enough? Great! KDE-4.1 is good enough? Probably, for like 20% kde users. Folderview is an adequate replacement for KDE-3 desktop? Bullshit.

    ReplyDelete
  11. @Muntyan: I'm not cheating. KDE 4.1 may not be good for *you*, but for others it may be well the case.
    Personally I have switched since 4.0, and while I wouldn't recommend a leap to 4.0 that wasn't well-thought, 4.1 (for me and also for other people) is mature enough.
    The moral is that while you can say that KDE 4.1 is not ready, such statements should beconfined to your personal experience, and not be generalized, since, of course, they are limited to your own experience.

    ReplyDelete
  12. Distros are switching to 4.1 because back when 4.0 was current, you promised that 4.1 would solve all the problems. Now it's 4.2 and I guess when 4.2 will be out and people still complain, it will be 4.3. :-/

    Distributions need to make their decisions on what to switch to well in advance, it's too late to tell us now "Oops, 4.1 isn't as good as we hoped, please wait for 4.2!"

    Part of the decision to ship 4.0 in Fedora 9 was that we knew that we would be able to ship 4.1 as an update soon afterwards (which we did, we now have 4.1.1) and people who really don't want 4.0 could just stay on Fedora 8 until that happens.

    That said, I disagree with your implied statement that 4.1 is not fit for consumption, there are some feature regressions compared to 3.5 which will be addressed in 4.2 (e.g. panel autohide, some Ark feature regressions), but mostly it's working. Heck, even the 4.0.3 release F9 was originally shipped with was essentially usable. It just had some annoying bugs (which all got fixed in 4.1 (or in some cases already in 4.0.4 or 4.0.5)), but the real 4.0.0 showstoppers were already fixed in what we shipped in Fedora 9.

    ReplyDelete
  13. PS: The "you" in my first sentence is a plural "you".

    ReplyDelete

Say something smart and be polite please!