Sunday, March 31, 2013

Debian: A SpeedBump on the Road to Innovation

by Dietrich Schmitz

I've watched the progress of Linux over quite some time and can't help but conclude that development in the Debian community has become dogmatic, stodgy, and held back for no other reason than pure politics and control issues.

There is nothing creative or innovative about that.  The Debian priesthood make their proclamations, anoint new members and continue to exert control of the pace of development with no interest in changing their software release management policy speed knob, currently set to: slow.  Slow as in 'sloth' slow.

As Debian prefers to resist change, change besets them.  You see, change is occurring on Internet time all around them in fact.  That's fast for those who watch trends in application development like me.  And the constant that is ever-present that Debian cannot ignore is that change will continue in overdrive while the Debian community sit on their collective hands satisfied by not doing what needs to be done.

Innovating on Internet time doesn't mean one has to lose control, nor does it mean that one will lose stability at the risk of frequent change.  It means that the pulse is being followed and striking while the iron is hot is imperative to reach profitability and bring product to market when and where it's needed--today and now.

So, life goes on, with or without Debian.  They have made their bed and sleep in it.  The need to politic continues unabated and the camp has their wagons drawn into a circle as if to prepare for attack.

It is fractious, unnecessary, and drives a wedge into the process of community sharing.  Sharing of ideas, sharing of resources is divided along political lines.  It is divisive and leads to internal turmoil all avoidable but the control must continue.  And it does.

Debian leadership will continue to apply their full will with impunity and seek comfort in their ability to exert only control, not innovation, not creativity, not sharing--just pragmatic politics--and continue to recede into irrelevance.  Happy Easter.

-- Dietrich
Enhanced by Zemanta


  1. First: good to see you being willing to speak not just about the good in the open source world but also about places where you think there are challenges or weaknesses. I see too much "advocacy" which is in denial of the challenges in the open source community (or sections of it).

    Second: I wish you had gone into a bit more detail here. I do not use Debian but I use some Debian based distros. In what way do you think Debian should be moving forward? How does their lack of progress affect the distros based on theirs? Would love to hear more about this in the comments here or as another post.

  2. I agree with this in the case of, say, Debian Squeeze. Using Sid, however, has revealed to me that Debian is quite capable of moving at a rapid pace, though at the cost of stability.

    If you don't like old versions of software, don't use Debian Stable (currently, Squeeze). Use another distro instead. That's the great thing about the choice we receive as Linux users: if we don't like something, there always exists a plethora of alternatives.

  3. Debian derivatives is like a cookie-cutter industry. It's become relatively easy to spin a distro, brand it and do little in the way of innovating. I am trying to move thinking off of a status quo mentality. Debian has been and is stagnating in its own policy making stupor and the real-world is passing them by.

  4. Slow food! Slow love! Slow Linux!

  5. jon_downfromthetreesMarch 31, 2013 at 5:43 PM

    I agree, but words like "profitability" and "bring product to market" are likely to have the Debian faithful casting spells.

    I've often wondered if there are considerably fewer people working on Debian than we are told. That might account for the interminable slowness.

    Sure, around the time a testing release is ordained as Stable, like now, things appear to be moving with some deliberate, but cautious, speed. But, a few years from now, when Stable users are running a version of Gnome released in 2012... well, there you go.

    If Linux, or any other software, required as much time to make itself reliable as Debian says it does, everyone else would have adopted their model. They have not, and the world's software, as imperfect as it is, is not falling apart every 5 minutes.

  6. I'll throw in my "5 cents Canadian" (hey, why not!)

    I agree that the parts of Debian that a 'typical user' sees are or can seem like they are stagnant (older versions of Firefox, LibreOffice, GNOME, KDE etc.) and this can be an issue for a lot of people, even if you attempt to get newer things via the Experimental packages.

    The back end stability though is a great thing - maybe not appealing to those who are into that, but for "Joe and Martha Six-pack" it would be okay - do they care what is underneath the UI, so long as it works? Probably not...

    So maybe a Debian branch that maintained the solid back end underpinnings but gave you the latest and greatest of Linux *user* software would be a winner.

    The problem is that many of the other distros out there (Ubuntu, Fedora etc.) give you the newest toys but the stability is or can be... questionable.

  7. Debian is an 'artifact' of the past. They are an illusion.
    Android and ChromeOS are real. They are TODAY and Now.

  8. You could say that, yes.

    Yet, as far as I know, you can't build Chrome OS on Chrome OS, and I don't think Android can build itself yet either (sure, you can build Android apps via AIDE on an Android device, but I don't think it can build itself directly).

    Google is using "Goobuntu" to do this primarily, I would assume (of course you could use Windows or OS X or any Linux) - and that is based on Ubuntu, which primarily consists of Debian Unstable packages.

    So, without the artifact of the past, there would be no today and now, let alone a future, via Android and Chrome OS :)

  9. Well yes, many Debian-based distributions are indeed forks/spins of Debian itself (or another derivative thereof). However, they exist to provide an *alternative* to Debian's bureaucratic morassity (if that's not a word yet, I'm inventing it :) ), rather than become part of it. Even if Debian itself moves relatively slowly (though this is quite debatable in the case of Debian Sid), it allows the Stable releases to be actually *stable*, rather than marketed as stable but actually rushed to meet a deadline. It's a slower process, sure, but sooner != better, at least not in all cases. Meanwhile, the arguably "better" spinoffs - Ubuntu, Mint, Knoppix, etc. - do what they can to differentiate themselves while building upon Debian's underlying design, using updated packages, novel user interfaces, unique optimizations, etc.

    It's also worth mentioning that Debian-based distros are not the only culprits of the hardly-a-fork phenomenon; the communities revolving around Fedora and OpenSUSE actually *encourage* a cookie-cutter-distro mentality with Fedora Spins and SUSE Studio, respectively.

  10. If Debian closes shop, a good chunk of the distrowatch top 100 goes down as well. That is like arguing that RHEL and CentOS are dinosaurs. There is a need for solid, stable, reliable, robust, trustworthy, tried and true OSs like Debian Stable. What's needed is a method to package Linux software so that uses of stable distros can use updated versions of user land applications.

  11. Debian user since 2000.March 31, 2013 at 9:06 PM

    Author obviously ignored the fact that Debian maintains software for 11 active architectures. Most 'derivatives' ship only i386/amd64 (32/64-bit PCs). Even Fedora/RHEL doesn't support that much hardware/software combinations as Debian. Take a look at and think twice about saying that Debian developers are 'slow' again.

  12. I gave it some thought. Sorry. Still feel the same--Debian developers are slow. In other news....

  13. Nice. Morassity. I am writing that down. ;)

  14. That's the second thing I've said that you've written down; I think a Linux Advocates Word of the Day is in order :P

  15. I can't help but wonder if the multi-platform support in and of itself is causing a significant portion of the developmental slowdown, almost as much as the morassity. I can't complain, though; it's nice having PowerPC support for my trusty eMac :)

  16. Indeed it is. I received it as payment for clearing passwords on some guy's Windows XP machines a few years back. Happily running Debian Squeeze now, and serves as a really noisy bedroom alarm clock (those speakers really pack a punch for their size), among other things.

  17. Debian user since 2000.April 1, 2013 at 12:11 AM

    You still don't get it. Debian is a public project, consisting of about 1000 volunteers, donating their free time and expertise, governed by a non-profit foundation. RedHat (sponsor of Fedora, owner of RHEL) and Canonical (owner of Ubuntu) are commercial entities, designed to make money and sell you their product (operating system, support and different services) as frequently as possible. You're trying to compare apples to oranges without second thought and complaining that they are not the same.

    Please take a good look at Linux distribution timeline at Wikipedia. See how many distributions are there, how many of them are Debian-dependent. It's easy to bash volunteer work for being 'slow', it's hard to get involved and help them.

  18. Here in the scientific computing arena, where stability is an absolute must, I will NEVER EVER use anything other than Debian or Centos. I will never repeat the mistake again of using distros like fedora or ubuntu where random things break down after doing a simple update and we simply can't afford to keep updating the OS every 6 months.

    Debian and Centos are made by power users for power users. REAL computing is done on them. Who gives a crap about internal politics when the end result is a rock solid Distro that rightfully earned the name "the universal operating system".

  19. Every linux disto that is cutting edge and forward looking is a step backwards. Gnome3 and Kde4 are both inferior to their predecessors, with Wayland kiss goodbye to Xorg.conf and its control. Ekiga was stable at the very outset as was Gnome Meetings in its hayday, but now it has waned into buggy instability. All good software are tools to do a job, many are excellent because of the simplicity of their initial design that gets lost along the passage of time. All good software comes in cycles, at their peak there is the Golden Age. For gnu-linux this has waned for not everything we call progress is progress. Personally, I have tried all the variations of linux, but to run the variety of programs I need, at the peculiar level of their development whilst not clashing with each other, I have only been able to easily achieve this with debian than any other....Take vsftpd and the madness of stopping users chroot, ...more secure....yes, practical, no...
    All the best Dietrich and Happy Easter!

  20. Some people actually want software that works flawlessly. With Debian GNU/Linux they get great software and can set up a system in seconds with APT. Debian runs the largest share of web servers for a reason. People want them to do their jobs day in and day out with no fuss. If you have more than a couple of dozen PCs and servers under your control, Debian is a huge advantage. One can control thousands of PCs without leaving your seat and upgrade OS and applications with a single command.

    I have been using Debian GNU/Linux a lot for five years and would not think of using any other general-purpose distro. The huge repository and APT makes it a powerful tool.

    The purpose of a distro like Debian is to bring FLOSS to the public ready to go. Debian does that exceedingly well. The purpose of such a distro is not to distribute flaky half-baked software from the bleeding edge. That's a waste of everyone's time. The purpose of a distro is to bring together a huge number of software packages in a way that they will work smoothly on your computer. That's impossible when distributing half-baked software.

    Ubuntu GNU/Linux is a distro considered by many closer to the bleeding edge and Canonical does develop some of the code in that distro but most comes from the FLOSS community and Debian's repositories. The first time I used Ubuntu GNU/Linux in an installation, Canonical had one package I installed to add some features clobber the configuration of another package which was working well so that none of 153 clients could boot. I don't get that from Debian because they have formal policies that one package does not modify the configuration of another. I have not installed Ubuntu GNU/Linux very often since then simply because I want the systems I set up to run reliably.

    Ubuntu and Debian testing have some similarity but I still prefer Debian's formality rather than the whims of Canonical. By using Debian, I have dodged disruptive changes like Unity, splashing local searches all over the web, broken things and threats to do away with X windowing system, one of the crown jewels of GNU/Linux wherever one has a mix of old and new clients and servers. Call Debian stodgy if you wish by love that it keeps my systems running flawlessly.

  21. When you're looking from a "desktop" perspective perhaps, yes. But keep in mind that GNU/Linux is also a very popular server OS, and Debian most certainly is among the more popular ones in that specific usage.

    Plus, as others have noticed, Debian is the foundation of many derivatives such as Ubuntu and MInd to name just two. For that reason alone the importance of Debian should not be underestimated.

  22. anon debian cowardApril 1, 2013 at 3:12 AM

    Two articles of yours today and they are intentionally against Debian and their derivatives. I hope at least you get well paid for the hatred.

  23. It's unlikely that an article on a site like this would really think that there is an equivalence between android and the Debian project, so one must conclude that the author is trying to be provocative to promote discussion.

    It is difficult to see how the way a group of people chooses to carry out an activity (in the this case the Debian developers) allows a different group a require then to do things differently. Conversely, it is difficult to see how the Debian developers' action in any way diminish others' abilities to act differently if they wish. So their right to do things the way they wish is entirely protected. One can therefore note or document the way they do things, but to assume they must manage the inevitability of change in only one way is not valid comment.

    Indeed, the very length of the Debian project demonstrates that they understand how to manage change in the IT world rather well. The many distros that have not lasted the distance shows that this is not an easy issue, and that reacting to everyone's views is not the way to survive.

    Debian portrays itself as "the universal operating system." It does this by continuing being excellent on servers by stability and using only proven innovations, by supporting a range of hardware platforms and by offering three positions with reference to any cutting edge of change. Assuming the author is discussing the desktop, anyone who knows Debian is unlikely to compare a "stable" desktop with a "testing" desktop or an "unstable" desktop. But all are Debian, which is what makes it such a valuable contribution to the Free Software community even if one chooses not to run it.

    Personally, I run Debian for servers by preference, and on older, lower-end laptops and desktops, where "testing" proves to be quite adequate. My life would be the poorer without it, and I believe that that value is partly to do with the way the Debian developers choose to run their project.

  24. Interesting. In my experience, Sid has been at least as stable as any other Linux distribution. "Stable" means that the packages don't change, not that it runs better. "Unstable", at the same time, means that the packages can change (get updated, added, removed, consolidated) at any time.

    To make Debian Sid very, very stable, just don't update it. :^)

    That said, there are times when Unstable is so unstable that it's best to wait a couple of months for things to settle down. Such as the change to clib6, or from Xfree86 to Xorg. But those kinds of transitions are very rare.

  25. Let me take the opposite side, and argue that the seeming "slowness" of Debian is actually beneficial for everyone.

    When Debian releases a Stable version, the goal is that every package in that release works, every time, right along side every other package. Localization, different architectures, different UI, all of it. What the Debian developers call a 'release critical bug' would be ignored by other distributions. Ubuntu is well known, for example, for their "every 6 months whether it works or not" cycle.

    By making this compatibility goal absolute, the Debian team gives a great incentive for the upstream application writers to do a better job. Making their applications 'good enough' doesn't make it, it has to actually BE 'good'. The patches and fixes are communicated upstream, and thereby become part of every other distribution as the software common to the entire world improves.

    Anyone who wants fast turnover can run Unstable.

    To complain that Debian is stale is to miss the forest for the trees. Debian stable is STABLE, for those who want rock solid stability.

    If you don't want that, don't use Stable.

  26. Your comment that Debian is not ChromeOS or Android shows that you are missing the point of what Debian is and maybe more importantly, what it is not. It is not meant to compete with ChromeOS or Android. It is not now, nor ever has tried to be the latest and greatest. It is supposed to be a full throated, free, stable, and intentionally vanilla operating system. There is a reason there are so many Debian based distros, why so many servers run Debian, and why people TRUST Debian.

    To say that Debian is an artifact because it's not ChromeOS or Android, is to say that ChromeOS and Android are just toys because no supercomputers or servers run Android or ChromeOS.

  27. Maybe I wasn't clear on that. My comment about ChromeOS and Android is that they have reached the commercial Market and truly represent innovation. Most people using both don't know they are using Linux. And it was done on a time scale of just two years. There's no equating them with Debian.

    There's too much time being spent acting sanctimonious and holy about what Debian represents. Like it or not, they are a speed bump.

    They won't help by being so and the competition is flowing around them on Internet time.

  28. Funny to see how you carefully prepend GNU/Linux to every reference. It's Linux.

  29. Oh I get it. It's all hype and a trumped up story. Sorry I don't buy it. Debian programmers have shown themselves to be lazy, slow and mean-spirited.

  30. The main problem with this theory is that Debian is the basis for nearly three-quarters of the currently active distributions. If nothing else, it is obviously a stable and reliable foundation.

  31. Funny how you think that is funny. I don't know any distros of FLOSS except Android/Linux and the *BSDs that don't use the GNU tools. Remember GCC?

    Of what are you an advocate? Division? FUD? Hate?

    Linux can't do much on its own. It needs tons of stuff besides the kernel to make a PC usable. A minimal bootable GNU/Linux OS from Debian is a bit over 100 packages, including a kernel. My working system has

    " dpkg --get-selections|wc

    2686 5372 73050

    2686 packages. It's not Linux. It's GNU/Linux. Calling the OS that I use "Linux" is like calling my car a straight 4 because that's it's engine. It also has batteries and electric motor-generators, a body and all kinds of gadgets too. The little woman would not have bought it if it was just an engine.

  32. Quite possibly, you may need to have that checked out. My guess is it's either hemorroids and/or a blocked bile duct.

    If I have raised questions about the 'status quo' and all of the sanctimonious hype that is floating around concerning your beloved Debian, I am sorry, but that doesn't mean I am creating, FUD, hate, or division.

    It means I see Linux not getting traction and simply moving sideways. And I question the motives of the Debian community for good reasons. Delaying or blocking a process from moving forward is in my book obstructionism and only exerting control, not innovating. Thus the term 'speedbump'. It doesn't stop innovation--it only slows it down. Debian is a speedbump.

    As long as simple exchanges (this) result in 'knee jerk' responses (yours) then it is quite symptomatic of just how poorly things are going. Try not to let your knee hit your chinny chin chin.

    And, please, be a good penguin, would you? ;)

  33. It doesn't really matter what you want it to be called, it isn't called "GNU/Linux", the community / end product has been branded "Linux" - so get over it.

    "Of what are you an advocate? Division? FUD? Hate?"

    Stop poisoning the community with crap like this, we don't need it.

  34. You are frustrated with Debian release police "Out when ready". Ok, stand in line with other 18yrs kids. You give arguments like one...
    For your info, Debian inovation is Stability, Number of packages and supported chip platforms...

  35. "The purpose of such a distro is not to distribute flaky half-baked software from the bleeding edge." Whilst your good self might be happy to rest on your laurels and accept old stagnant software, some of us enjoy the fact that there are new innovative _bleeding edge_ projects out there. I wonder why archlinux the bleeding edge distro is most commonly used by developers & power users, sure its not for everyone but these type of communitys actually push things forward. Debian too was bleeding edge 20 years ago.

  36. Archlinux too is community driven and run by a small group of volounteers, they manage to give us the most modern software, granted it may or may not be as stable as debian but my point is it matters none if they are sponsered or not.

  37. Considering my last posts on the subject, this is going to sound funny coming from me. However, Pogson is right: the proper name of Debian is "Debian GNU/Linux."

    He is wrong, though, to use "Ubuntu GNU/Linux." That distribution has never used "GNU/Linux," and these days usually simply uses "Ubuntu."

  38. Yes and I insist we respect Otto the inventor of the 4-cycle combustion engine by referring properly to the OTTO/Mobile, not car, not auto. It's just wrong! ;)

    try to search on 'gnu/linux' on wikipedia, see what page it returns:


  39. What about distros that use busybox? Should we call them busybox/Linux? Should I call Slackware KDE/Linux because it happens to have KDE installed?

    The GNU userland is nice to have, but it's not the one defining characteristic of the operating system as a whole - or even the main one - and to give more weight to the GNU components than to X components or desktop environment components is to wallow in inaccuracy and absurdity. If the distro consists *primarily* of GNU components, then fine, but very few of the more common distros are limited to that.

    Regardless, however, I think we as a community have *far* bigger things to worry about than trivial semantics - like hardware support, as well as establishing "Linux" as a household name - if we want to bring the free software movement from obscurity to the mainstream.

  40. As I wrote recently, I've given up using "GNU/Linux." But for me it's only common courtesy to refer to "Debian GNU/LInux," because that's the name that the project actually uses for itself.

    In the same way, it's rude to use "Ubuntu GNU/Linux," because Ubuntu has never called itself that.

    This has only indirectly to do the GNU Project, and everything to do directly with respecting the wishes of the members of a project.

  41. Scroll down a bit and read Ryan Northup's response.
    Me: it borders on absurdity. We have much bigger fish to fry.
    Thanks Bruce. D

    "What about distros that use busybox? Should we call them busybox/Linux? Should I call Slackware KDE/Linux because it happens to have KDE installed?

    The GNU userland is nice to have, but it's not the one defining characteristic of the operating system as a whole - or even the main one - and to give more weight to the GNU components than to X components or desktop environment components is to wallow in inaccuracy and absurdity. If the distro consists *primarily* of GNU components, then fine, but very few of the more common distros are limited to that.

    Regardless, however, I think we as a community have *far* bigger things to worry about than trivial semantics - like hardware support, as well as establishing "Linux" as a household name - if we want to bring the free software movement from obscurity to the mainstream."

  42. Dietrich,

    I've been debating whether or not to comment on this post of yours since as you may have guest... I love Debian and I use it every day (even at work). To me it has proven its worth as a distro that gets the job done without having to have the latest and greatest. It's a workhorse distro that has plenty of innovation under the hood to last for the next decade too.

    What I don't see in this blog post of yours is evidence that Debian is a thing of the past.

    Was there a recent interaction of yours with a group of Debian developers that didn't go well? Were your contributions to a Debian based project not accepted?

    One thing that is also overlooked when discussions like this one take place is that
    Debian is one of the few distros that has polices in place to ensure that things work as people expect them.

    Things might move slower than let's say Ubuntu because there is no hurry to reach a tight deadline of 6 months. It takes two years to produce the stable branch of Debian and with good reason.

    In the open source world, speed kills. It makes for really bad design decisions,
    incomplete implementations and patched up code to the point of being un-maintainable thus wasting precious man hours and producing only abandoned unfinished unmanagable projects.

    Debian also realizes exactly what you described which is why it has exposed the world to the unstable and testing branches of the distro (instead of keeping them internal). More eyes to look at the code and fix bugs. Just take a look at their security advisories. You know they take their contributions seriously.

    Debian has been powering everything from Nokia phones and ARM based systems of all sizes to production servers of all sizes. The first live distro,
    Knoppix, is a Debian-based derivative.

    If you still don't have enough data that Debian is not a speed bump but a cautious carrier of innovation then I think I better stop here.

    With all due respect and with no intent of being nasty, hostile, rude or insensitive I just don't understand what or who from Debian pissed you off enough to write a hostile article like that.

  43. For desktop Linux, users want the latest visual effects, the most recent browser, the newest media players, the latest version of their preferred IDE and programming language, etc... etc... For servers, stable and secure trumps all other concerns. Debian's rate of change makes it an outstanding server distribution and a poor desktop distribution. There's nothing wrong with that combination.

  44. I have challenged the 'status quo' mindset on this. Stable Debian may be, but the release management policy is deficient. Thus the term speedbump.

  45. GNU_Prometheus_LeagueApril 14, 2013 at 1:07 AM

    Yeah, and it doesn't matter how misguided or euphemistic a product or concept name is because we just have to accept it and "get over it." So we're stuck with the media spin master's destruction of our words, such as turning the honorific term hacker into a synonym for criminal; euphemisms like "pacification" for aggressive war crimes, or "Correctional facility" instead of jail, Sanitary Engineer for trash man; oxymoron's such as Apple tech support, Business Ethics, cheap gas or Corporate Culture.

    Taking your point of view would only serve to paper over the moral/ethical/political nature of the Free Software movement. We GNU/Linux Free Software advocates are about bringing freedom to computing, not merely putting out products for corporation. Focusing attention on this fact is not "poisoning the community with crap." Its bringing attention back to the very essence of what we are about -- freedom.

    So contrary to your assumption language and societies are malleable and we reshape language to focus attention on the important social justic issues at hand.

  46. Perhaps, but even in a philosophical context, the FSF has its roots in 60's/70's hacker culture, which encouraged openness and the ability to freely collaborate, as do the BSD folks. Software freedom is not unique to GNU, nor did the GNU project originate that concept; said concept had been prevalent - hell, even close to mainstream - even before Unix existed at all. Stallman only formalized an existing mentality.