05 May 2008

Why should I "fight" for desktop linux? or Has Red Hat betrayed me?

Sorry for the delay, this happens when the feedback is happening elsewhere , clashing with May Day and a weekend. (note: I expected a bit of controversy after the Desktop piece in my webcomic and was wondering about the silence).
As usual, Radu is strong and direct, he says "let's be honest: this is what a fanboy does when he doesn't want to admit the betrayal" (linking to the already famous official position from the Red Hat Desktop Team) but he also ask a question which I think deserves an answer:

"By the way, can you explain why should you fight for "Linux on the desktop", when Red Hat Inc. has stopped doing it?"

And the answer is simple as that: Because *I want* to run Linux on *my* desktop.

And, from what I know, there is a large number of other Fedora contributors who think the same, they want to run Fedora on their desktops and they want a first class experience doing so.
Of course, there is also a large number of Fedora contributors who are content (some are even happy) with running Windows, OS X or that other Linux distro on their desktop (even I know some) and that is their option.

Here is a short Q&A session expanding on the short answer above:

I am disappointed by the Red Hat's position about the desktop?
Sure. I use so-called "traditional desktop". I see the pluses and the minuses of Fedora in this role and understand the need of a large amount of work needed to improve it. And the need of a strong player (and $$$) behind this work.

I feel betrayed by Red Hat in this matter?
The (probably unexpected) answer is: No. You feel betrayed when you have expectations from some entity and see those expectations vanishing in the air. Well, at least in the last 5 years, Red Hat never talked about a frontal assault on the traditional desktop and all the official declarations were against it.

Do I think Red Hat is wrong with that position?
Yes. I think it is short sighted and in a few years will come back and bite them in the ass. But I also think I understand their reasons and can't provide a better solution (and isn't my business of finding one).

Should I care what happens with Red Hat?
To some degree: Yes. I am a Fedora user and a Fedora contributor, Fedora and RHEL are different distros, with different goals and audience. But Red Hat and Fedora are in a symbiotic relation, they will fail or succeed together. And I invested work, time, resources and emotions into Fedora.

I may be completely wrong in my analysis, but I think it is all about bang for the buck, spending your money in an efficient manner. And currently the best return of your investments in the Linux world seems to be on the server side.

Years ago when Red Hat choose to focus on RHEL and changed the old RHL into the Fedora "community", it was a huge outcry, betrayal accusations and flamewars on the entire Linux world. But so far it proved as the correct decision: it created a healthy company growth with a viable business model, which allowed them to expand and hire a lot of great hackers (and not only hackers) to work on various areas, including the desktop.

Desktop is hard, it is a money pit: Corel tried and failed, Linspire tried and failed, Xandros tried and failed, Novell is trying and does not look like succeeding, Ubuntu is still working as a charity and does not appear (so far, but it will come) to try and sustain itself.

Now Red Hat looks like is using the same strategy and focus on JBoss. Money invested there are expected to produce a healthy revenue stream which would allow them to hire even more hackers to work on wonderful things, including the desktop. And we have to take that and be happy with it, even if we wished to see Lunix commercials on TV or so.

Now to the "short sighted" part. While the move to the RHEL/Fedora combo may have proved as a success in terms of revenue, at the same time it acted as one of the primary factors contributing at the creation and success of its primary competition in the Linux world (despite all the collaboration in development, nice talks and Kumbaya singing, the distros are in competition).

And these days, not pushing the desktop is strengthening the competition. Desktop Linux is happening and it is happening elsewhere. And it is important for the server world because it builds mindshare. The desktop users of today are the decision makers of tomorrow - they will decide what to deploy in their corporations, and that most likely will be the server version of the distro they know from their desktop.

From my external point of view (which I acknowledged may be completely wrong) I can see an additional reason: corporate culture. Red Hat, as a corporation, seems to understand the server world well (as shown by the feedback from their customers) but not understand so well the consumer desktop (looking at some RH desktop initiatives: OLPC is a failure, going to Windows, RHGD has silently died, GNOME Online Desktop is a few RPMs in F9, but nothing in the Feature List, Mugshot is in the same boat as GOD). Usually is a god thing to stay with what you know and do it as best as you can, and Red Hat may be doing just that.


  1. You're making the common mistake of thinking Red Hat has in fact stopped working on a desktop. This clearly isn't the case because there is a lot of work from Red Hat going into Gnome for both Fedora and otherwise. We spend more work on the desktop than that other company does that claims they own the deskop, I'd wager.

    So, before you make a long blog post about press releases, I'd recommend going back and figuring out what is really going on. What is going on is that Red Hat doesn't have a desktop-specific relese... but wait, what's Fedora? Life is still good, and I think you know that.

    If anything is short sighted, it's that they don't make their rather large contributions to the desktop arena more obvious.

  2. I don't think I make that mistake, I said "which allowed them to expand and hire a lot of great hackers (and not only hackers) to work on various areas, including the desktop".

    What's Fedora? "a Linux-based operating system that showcases the latest in free and open source software". It is a general-purpose OS, not a desktop OS.

  3. > OLPC is a failure

    OK, why isn't the management sacked for that error? And why everybody told me I am an idiot when I said that this is a waste of time for RH? (I said it for free, the management was paid for that.)

    > RHGD has silently died

    Once again, why isn't the management sacked for that error? When I called it vaporware, a secrecy and a hoax, I was told I am stupid. Now where is the truth?

    > GNOME Online Desktop is a few RPMs in F9

    I won't say it again... but you know what I mean...

    PLUS: no matter what they *do*, what they *said* is catastrophic from the PR standpoint. People will understand that RH doesn't care if they use Ubuntu or Windows!

    Someone should have paid for that.

  4. Why do you ask me? I don't know... Maybe the success criteria for those projects wasn't to get a deliverable product... maybe they achieved their goals or maybe some personnel changes/departures at RH (not at the very top) are related to that...

  5. @beranger-org

    If you are a stockholder with a large enough share (votes) in the company you can always bring it up to the board etc on why people were not sacked. And I am guessing you mean publically sacked with an a ceremonial or literal burning to make sure the statement of contrition and failure are known. Ah the good old days when we could just kill failures and stake them out in the open to remind people why they should never try something that they are not guaranteed to succeed in . Should have done that with that Linus chap when 0.04 didn't compile.

    In most cases, people are sacked for stupid mistakes... you just can't say it because of lawsuits. In the case of a making an educated gamble and failing, they are usually off to find something else.

    @nicu.. as much as I would like to see Fedora as a general use OS.. I can't use it for that because most of my enterprise and general usages require system lifetimes of 3-4 years minimum. As a desktop with new eye candy every 6 months... I can and do use Fedora. So from my enterprise view.. it is a desktop os.

  6. 1) RHGD isn't dead. You'll find that when creating buisness that will last longer than a fad (read sustainable) things take longer than the speed of just throwing out whatever you've currently got and hoping it sticks.

    2) Red Hat still does more desktop development than most other distros. I'm even elected on the GNOME Foundation board and allowed to spend my work time on board issues.

    3) If you think you have succeeded at everything you are just picking easy problems and not learning from your failures. Or more likely you are picking easy problems, still failing and deluding yourself.

    4) Red Hat still works on Online Desktop. Developments hard - help out if you feel it isn't moving fast enough and keep an eye out for Havoc. I have no idea what his company is doing but I have to assume it has something to do with devices, Linux and the internet. It is fine if he takes that market because that is just growing Linux (it isn't a zero sum game here and smaller companies tend to move faster in emerging technology).

    5) Monopolistic, entrenched markets suck, we need to route around it instead of using in the box thinking - this is why the year of the Linux desktop never came. The key phrase is "Traditional Desktop", now go back and read the position paper.

    So I would say that I understand nicu's concern and thank you for the praise but your post was a little more downtrodden than the way I see things. Words are a strange thing, they are very specific but people tend to generalize them and any words hitting too close to home - too close to our hopes and dreams - can instantly be transformed into black and white - good and bad when really it is a grey area. A seeming negative today can mean double the payout down the road.

    Note this is all my take on things and do not reflect my employer's own thoughts on the subject.

  7. @J5: I appreciate your long and insightful reply.

    1) I can't find the URL, but I remember reading a recent interview with a high ranked RH executive (a VP?) much to that effect, my opinion wasn't based only on the position paper.
    Anyway, can we expect a release of RHGD this year?

    2) Of course, you and a lot of your colleagues are absolutely kicking ass. And I use Fedora on the desktop every day. But a "desktop distro" should be something more.

    3) Sure, I can list some successes, for example I think Bluecurve was a success at its time.

    4) In F8 GOD looked like a priority, which isn't true anymore (and IIRC it was supposed to be the default mode in the Global Desktop). And Havoc looked like the main person behind GOD (and its visionary). This make one wonder about a possible link between GOD development and Havoc leaving RH (I don't expect a straight answer here).

    5) But I think the "customers" are not ready for a non-traditional desktop, and this may be a lesson to learn from OLPC. Mr. Negroponte says he was asked by potential "buyers" if the device can run Windows. Those people in fact asked for a traditional desktop.

  8. @J5:
    > The key phrase is "Traditional Desktop"

    What is a "traditional desktop"?!?!

    Linux has X since ages. X has window managers since ages. There is a Linux desktop since ages.

    What we don't have very clear yet is the focus on the "checklist of a modern laptop/desktop" user. However, I had device mounting way before HAL was born, I had reliable hibernation (on the hardware I had at the time) with FC5, etc. WHAT PREVENTS RH TO GO ON?

    The excuse of the "monopolistic etc" is a lame excuse. I personally don't believe in such a thing as "the year of the Linux desktop", however the PH public statement was wrong. I DON'T NEED TO READ IT AGAIN: ALL THE PRESS IS SAYING 'RH HAS DROPPED THE DESKTOP'. If you (you===RH) wanted to say something else, you should have written something else.

    > Note this is all my take on things and do not reflect my employer's own thoughts on the subject.

    When you are hired at RH, are you brainwashed to end all your public expressions with that disclaimer?

    "-- Do you want fries with this?
    -- Yes, please, but note that this doesn't reflect my employer's views on the subject."

  9. "Traditional Desktop" is usually understood (and I used this notion in my piece) a Windows drop-in replacement. something that a regular Windows user would use without changing his workflow.

    The monopolistic excuse is true to a large degree, right now I just got trolled on YouTube about a GIMP screencast, being labelled as "photoshop rip-off". And don't get me started about someone who is pestering me about voice chatting over YM in the last days (in a half-business matter), thing I can't do with my Linux desktop... Really I hate to reboot just for that stupid thing.

    I find his disclaimer adequate, otherwise I could take his words and use them as the official RH position, which isn't the case. Without the disclaimer, he would have to post anonymously or maybe not post at all. I prefer with the disclaimer.

    As for "WHAT PREVENTS RH TO GO ON?" may I make a speculation? Maybe RH does not want to compete directly with Microsoft, nobody competed so far and won (not even Google, their battle is still going on). Or maybe there are not enough money in this market.

  10. Nicu,

    Try gyachi which I maintain in Fedora. It supports two way voice and video chat.

  11. Yeah, you are right about gyachi, I knew about it, even tried it once after I saw your first blog post about it but couldn't stand its interface...

    I wonder if is possible (I expect not) to run at the same time Pidgin for "normal" IM and gyachi for voice... maybe just fire gyachi when needed and reconect Pidgin after, I guess is better than rebooting.

  12. Just tried gyachi for voice. Quote for the dialog window:

    cannot run gyvoice due to the mising files:

    Neat stuff... (yes those files can be found easily)

  13. The interface is horrendous but very functional and stable. What you can do is run the webcam/voice chat in another id. Let's say nicu_vv and use your regular id in pidgin for actual conversations.

    Another thing to try is the very latest version of empathy where telepathy has proper native support.

  14. Don't listen to beranger. He is one of those people who just complains but never actually wants to help out. He's going to flame me for this but hey he is the one who said on his own blog that he doesn't need facts to state his views.

    On to Nicu's questions:

    1) I'm not sure when RHGB is coming out since I am not on the desktop team anymore but rather on the Fedora team. Right now I work on streamlining process, part of which will hopefully translate into a better developer desktop and eventually translate to the consumer space for Fedora/GNOME. Online Desktop was never part of the initial plan for RHGD. Note that OD is very much a work in progress that slowly gets integrated as parts stabilize. It isn't going to be some big bang thing. If you look at any other development process which looks like it just appeared you just didn't see the sausage being made, only the final product. Open Source doesn't have the luxury of silent development time so what you have today sort of just gets ingrained as how things are so the pace seems pretty slow sometimes. In my own experience, did you know D-Bus is something like 6 years old. It took two years for us to actually start using it and another two years when I started working on it to get it to 1.0 and widely deployed. Not to mention the time it is taking to get all the bits above it right and that is all just plumbing so we can do the real interesting work. If I told you that is actually fast in terms of development time for major components you might not believe me. As in any project it might get scrapped to focus on something else (like the box set was for the Fedora/RHEL split which worked out great by actually forming a community), you can never predict those things but it is done with a lot of analysis and isn't just gut feeling.

    2) Which is where the fedora community should step up. Just because Red Hat isn't creating a product doesn't mean Fedora is not a Desktop distro. Again we do set a lot of the direction for the desktop but we do that upstream where it should be. If that work gets projected as work of upstream developers instead of Fedora developers well that how it should be perceived. Remember we are fighting for the Linux Desktop not just the Fedora desktop. The market is just too small for us to be infighting while MS laughs all the way to the bank.

    3) I would say it was a success and a failure. It was a huge step in getting Linux to look more modern (I know now it looks dated but that is the consequence of subjective things). The failure is it took a lot of effort to maintain. You know that with the Echo theme which is still not done. But we learned a lot in the process and the reward was worth it.

    4) I think it was just announced as one of the ways we could modernise the desktop but it is like a car show. You'll see prototypes which kick ass but not ready to sell along side the admittedly less sexy stable models with incremental improvements. You can run the online desktop but it is such a paradigm shift that unless you productise it or get it to a point where people are comfortable with it people will just try it and then go back to what they are used to. This is the problem with chasing the traditional desktop (it is one of the problem MS has with Vista). If you think the Mac's sell well because they are better desktops you've missed the point that they sell well mainly because Apple came out with a product called the iPod in a market segment that was not monopolized. If you want to see the real catalyst to Mac sales it is the iPhone and the iPod. The Mac is simply a better way to fill those devices up with songs. Design has always been the strength of Apple but still they only have %5 of the desktop market after years of obscurity and at one point being propped up by MS. They played to their strengths, focused and thrived.

    This is what Red Hat is doing. Focusing on their strengths. You talk about Havoc. I don't know much other than what the paper said, Red Hat is not focusing on consumer products. I have to believe that is what he wanted to do so now he has this company Litl in which he can concentrate on that segment. Remember that the paper noted the Fedora desktop so while we don't have products Fedora has desktop projects. The direction of Fedora is not tied to the Direction of Red Hat other than we provide resources in the areas we care about. That is what the spins are all about, the ability for people to craft Fedora into what they feel they need.

    4) We are still just performing arm chair analyse here. Nothing has actually come of that yet.

  15. 1) Maybe the initial plan wasn't to link the two, but this is how people read the initial announcement, here is one sample from the press a year ago.
    I understand, and probably a lot of people understand that developing takes time (even if the time is longer than they hopped). But you did the right thing with D-Bus and talked about your work and the development status, so people knew it is alive and improving, which is not the case for GOD, lately all I hear from it is from rare posts, mostly from mpg, on the fedora-desktop-list. And surely is not the case for the RHGD, which was claimed to be ready to run by RH executives back in October, complete with implementation details (RHEL 5.1, GNOMe 2.16, etc.)

    2) There is one thing I would like to see about the Fedora desktop and this is a vision. Today our desktop looks more like a pile of cutting edge technologies thrown together, engineered brilliantly but not polished enough.
    And I think such a vision should come from someone who is both known (respected, trusted) and capable to produce a god vision. But I am not sure we have such a person.
    For example after bclark's departure, I'm afraid there is no professional designer remaining on RH/Fedora.

    3) I call Bluecurve a success, even if we happily got rid of it in Fedora in favour of Nodoka/Mist because with the help of Fedora we learned a lot about the need of visual consistency. For example Tango can be seen in a way like a successor of Bluecurve.

    4) At the time, after reading about GOD I tried it and decided: it isn't for me. But I liked one thing about it: it had a vision, even if, from my point of view, not the right vision for me.