Slashdot Mirror


Open Source In the Datacenter: It Was Never About Innovation

An anonymous reader writes "The secret to open source innovation, and the reason for its triumphal success, has nothing to do with the desire to innovate. It's because of the four freedoms and the level playing field (and agility) that was the end result. It's like Douglas Adams' definition of flying: you don't try to fly, you throw yourself at the ground and miss. This article explains why it was never about innovation — it was always about freedom. Quoting: 'When the forces of economics put constant downward price pressure on software, developers look for other ways to derive income. Given the choice between simply submitting to economic forces and releasing no-cost software in proprietary form, developers found open source models to be a much better deal. Some of us didn't necessarily like the mechanics of those models, which included dual licensing and using copyleft as a means of collecting ransom, but it was a model in which developers could thrive.'"

20 of 100 comments (clear)

  1. Most Software Is Shit by Anonymous Coward · · Score: 4, Interesting

    90% of everything is crap, but at least with open source you can find out why instead of waiting for the developers who can't reproduce your problem.

    1. Re:Most Software Is Shit by houstonbofh · · Score: 5, Insightful

      90% of everything is crap, but at least with open source you can find out why instead of waiting for the developers who can't reproduce your problem.

      Don't forget a total lack of license management, the purgatory of IT. Essentially, with Open Source, you can spend less time dealing with how to get the software, and more time working on interesting stuff.

    2. Re:Most Software Is Shit by epyT-R · · Score: 3, Insightful

      It's not even about just getting the software, it's about preempting lawsuits. Better to just go GPL/BSD/PD since they are easier to comply with.

    3. Re:Most Software Is Shit by dgatwood · · Score: 2

      Unless it's AGPL, in which case it does, for most normal people's definition of "using".

      --

      Check out my sci-fi/humor trilogy at PatriotsBooks.

  2. this is "news"? by murdocj · · Score: 2

    This is just an opinion piece, not even remotely news.

    1. Re:this is "news"? by drinkypoo · · Score: 3, Insightful

      This is just an opinion piece, not even remotely news.

      And is it "stuff that matters"? Not to me it isn't.

      You fellers (or whatever) are stuck in the past. The closest thing to a motto on the front page now is at the bottom, and it says Slashdot is a Dice Holdings, Inc. service.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    2. Re:this is "news"? by 93+Escort+Wagon · · Score: 2

      Yeah, the article mixes a couple different subjects that don't really have a whole lot to do with each other much of the time - development and datacenters. Most developers aren't doing anything remotely relevant to the datacenter.

      Open source works in the datacenter because it's cheap, relatively easy to manage, and because tools are available that let it scale up fairly easily.

      And while there are successful, large projects that are open source... it's harder to see the argument that open source is the tool of choice for developers - at least those who are trying to make a living at it. Maybe if you limit the scope to hobbyists or side projects...

      --
      #DeleteChrome
    3. Re:this is "news"? by mjwalshe · · Score: 2

      You can prototype in python with numpy which is what a lot of people do nowadays.

  3. Innovation rarely exists. by jellomizer · · Score: 4, Insightful

    When people say Innovative, we think of something that when we see it, we go Wow this is so cool I would never think of of that myself, and usually throws the rest of the industry in catch up mode.

    Now the iPhone (not the iPad) was an innovative idea. Phones before the iPhone had external keyboards, at the expense of of screen size, or thickness. The idea of very few real buttons at the time was very foreign to us. And using gestures seemed almost impossible, as many early gesture systems had a lot of complicated gestures to get tasks done.
    The iPhone wasn't innovative based on its features, there were other companies that had phones with more features or better hardware. But the innovation was able to successfully make a phone, that the advance feature were accessible and to the end users. The idea of say browsing the web on your phone, or have it as your main method to check for email seemed silly before, today it is quite common.
    What happened after the iPhone kicked off, it threw the Industry in catch up mode. It took years for good Android phones to get into the market to start competing, and these new phones all are based on the iPhone.

    Now the iPad isn't that innovative, it was easy to realize you take your iPhone and just give it a bigger screen, and fit better processing.

    Other innovative products.
    ID software 3D shooter. Wolfinstine 3d and Doom. They had some wire-frame attempts, and a few polygon based games. But games before that for the most part where 2d sprite based (Side Platform like Mario, or top down like Zelda), specificity for fast paced action games.

    Nintendo Entertainment System. Unlike the Atari and other predecessors it didn't give any allusion that it was a person computer, just a straight game console. Priced more affordable than the others, and focusing on games.

    Innovation is very rare. Most of the time it is copying someone else idea and tweaking it so there are different set of trade offs. Now their tweaks may change the market, but not as much as a innovative product.

    How you choose to license your product, isn't really that big of a deal. Open Source, sure people can tinker with it coming with some new ideas. Commercial Software will have paid employees trying to come up with something new.

    --
    If something is so important that you feel the need to post it on the internet... It probably isn't that important.
    1. Re:Innovation rarely exists. by Anonymous Coward · · Score: 5, Interesting

      In my opinion, the problem is the misdefinition of the word "innovation".

      An innovation is something new. Something revolutionary that changes everything.
      The transistor was an innovation. Rounded corners and ultra-thin device form factors are not.

      In the world of free software, tools, apps, and other software constructions don't overtly aspire for this goal. They just want to get X task completed. Most tools and products start as ugly assed hacks, thrown together with haste.

      The magical alchemy of free software, is that if somebody else expends the energy making that ugly hack, and shares it for free, others can snag up that hack, look under the hood, and either use it as-is, or use the energy they would have used to create their own ugly hack to beautify and refine the hack they just found, and make it better for servicing the unique twists of that person's requirements when doing that task. This could be anything from adding new features, to fixing dirty code work and inefficiencies in the logic. Des not matter. The effect is the same.

      Over time, the dirty hack becomes something the original author never envisioned, but increasingly more innovative, as more people look at it, and add clever improvements. It does not come into the world to change anything, just to do a job.

      It is an organic, evolutionary process. Something "barely fit" for the function undergoes selective pressure, and unrestricted replication, and intelligently guided evolution. The latter part is why it reaches "innovative" local maxima solutions to problems quickly.

      Trying to upset the applecart, just to upset the apple cart and change the world is a monumentally difficult task to "just do". FOSS does this effortlessly, one clever hack at a time. The payment the innovators receive in return, is better employment of their time (for the few minutes or hours of time they invest each, they all get demonstrably better software than they could have produced from scratch in that period of time, and if they improve the software and release under the license terms, then the time they spent adds value to the next person in the chain. It isn't about monetary compensation; it's about time use.)

      Proprietary software tries to leverage the time and energy of a small group of talented people, to prduce a product of greater sophistication than an individual software hack can produce in a sensible amount of time, and extort money out of them for the service of providing an already made package that should suit thier needs. (Should). This is done to get a slightly higher amount of monetary valuation of "time" from the customer, and offer a "bargain" in time expenditure vs value to the customer.

      (The software company pays their employees a certain financial compensation per hour worked, which is summed to help arrive at a production cost figure for the product. The proprietary vendor then amortizes that cost over an estimated userbase, and arrives at an MSRP, and from there a transaction for the finished product can be conducted. The msrp is higher than the amortized cost per unit, the price of the product for the customer is considerably lower than the valuated figure for the time it would have taken them to mae the product themselves. Both walk away with value.)

      With foss, this methodology is disrupted; there is no money seeking middle man. The value added by each small successive evolutionary step improves the software. They get the benefit of many times this investment, the longer the product stays in active development. Linux Kernel alone constitutes millions of man hours of coding time. If you spend 1 hour making a small improvement to current trunk, and have it accepted, you still have over 1,000,000:1 value return on the time. The next person gets 1,000,001:1 return. Etc.

      This feedback allows foss to grow and evolve radically faster than proprietary software could ever hope to achieve, especially as the development life of the product increases. Proprietary software has recurring costs in deveopment. FOSS has recurring returns.

      FOSS is a true innovation in software development.
      Thin screens and gestures pale in comparison.

  4. Re:CopyLeft ransom? by c0lo · · Score: 2

    Could someone explain how CopyLeft ransom works?

    I'll explain to you how CopyLeft works if you pay me.
    If you pay year-after-year, I'll keep you updated with how it works over the time and... I'll even allow you to call me twice per year.

    --
    Questions raise, answers kill. Raise questions to stay alive.
  5. GNU GPL FTW by Anonymous Coward · · Score: 5, Interesting

    When I write code for personal reasons, I always release it under the Affero GPL v3+.
    It saves me the time and effort of attempting to monetize or control every little snippet of code that I write just for fun or just to learn something.
    It also ensures that nobody can commercially exploit the code without A) paying me for a non-GPL license, or B) contributing back to the community.

    As a side effect, it makes a great way to show off my coding skills to potential employers.
    They can look me up on GitHub and evaluate my code and skills, but they still have to pay to play.

    I'm not a libre software zealot. I don't believe that everyone is under a moral obligation to release their source code.
    However, I do find the Affero GPL effective at protecting my non-commercial interests and providing an assist on my commercial interests.
    That is why I use the license, and encourage other software developers to do the same.

    1. Re:GNU GPL FTW by unixisc · · Score: 2

      How exactly does that work? You provide a package under AGPL3. I pick it up & use it, and run it - maybe on your server, maybe on mine. Everything I do is internal facing - I put it on an Intranet, but not the internet As a result, nobody other than me & my colleagues get to use it, I don't add a thing to it, so your source code is available to anyone who wants to see it - my colleagues, while they run it on as many computers as needed. Instead of paying for a closed source package, I got yours for free, and use it. I don't contribute any code 'back' to the community, since I don't write it, and since I'm not distributing it, your AGPL license works just fine.

      AGPL - the way I understand it - is that if I provide a software, but hosted on my server, not downloaded, I have to provide the source code to anybody who runs it, if it is on my server. If they then take that source code, add things and put it on their server, they have to provide the source code. However, your model doesn't touch those who use your service - if there is one - for purely personal consumption, and neither redistribute in the form of code, nor as a service. Such people would be fine, whether it's under GPL/AGPL/LGPL w/o paying you for anything, unless you charged them for the software in the first place (but they can still get it for free from a third person)

  6. Why FOSS? by unixisc · · Score: 3, Interesting

    One thing I don't get - if there is a downward pressure on prices on developers, how does adapting an Open Source model help them? It's not like they get extra money for it if they reveal their source code.

    Also, the 'four freedoms' have never been about making better software, as RMS never tires of pointing out (and it shows). They've been an end in itself. If you write a software - no matter how bad, but simply put it under a A/L/GPL license, RMS would be pleased. Your software respects the 'freedom' of your neighbors, who you must help, as per Freedom 2.

    But I doubt that the desire to put Open Source in the datacenter had anything to do with any 'freedom'. It was about putting better software out there. Since the existing datacenter hardware was tied to the support contracts that a Microsoft or Sun/Oracle or HP would provide, moving to FOSS meant that any datacenter that adapted it would determine its own support timelines, since the open source meant that they could hire their own developers to maintain it beyond upstream support, and also, the upstream projects had no strong reason to EOL a version, unlike commercial entities.

    The innovation part - this part is not completely true about FOSS, since there ain't millions of programmers interested in the project, and so the software usually doesn't get examined except by its developers, and maybe some very interested customers. Where FOSS helps is that if a customer has esoteric hardware, the software can usually be ported to it to exact the maximum life out of the system, as well as provide a uniform software platform for heterogenous computing environments.

    1. Re:Why FOSS? by drinkypoo · · Score: 2

      One thing I don't get - if there is a downward pressure on prices on developers, how does adapting an Open Source model help them? It's not like they get extra money for it if they reveal their source code.

      That suggests (maybe even begs) the question, can they get extra money for it if they reveal their source code? You always reveal your source code to your employer in a "traditional" programmer-getting-paid relationship involving corporations (or at least companies) and groups of programmers, marketers, et cetera.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
  7. Re:Everyone wants something for free by GerryGilmore · · Score: 4, Interesting

    Close. In my case, as head of technology and development at a small outfit then using SCO Unix, it was a combination of factors. First, and most important, was gaining some level of control of the underlying software stack. A couple of examples: We installed the SMP package on a customer's system. Random crashes and panics became too common. We replaced the server - no joy. Having a support agreement with SCO ($$$), we called them for assistance and their response was "re-install the SMP package". When I explained that we'd already done that, they said "well, do it again". Another time, we needed their DDE-RPC package to run some CSTA software. When I tried to buy a copy, they said "nope, we discontinued that package". I offered several options: we'll pay for it, but not ask them for support, etc. No, no and no. It was about this time one of my techs who'd been singing the Linux song finally handed me the pack of Yggdrasil floppies and once I finally got it loaded and started looking at the source code for *everything*: kernel, compiler, utilities, etc. my jaw hit the floor and I knew that the world had shifted forever. We started then on a migration project - which took a couple of years - and we've never looked back. Worth every penny that we didn't pay to SCO, but did pay to our engineers.

  8. Re:Gnu.org = biased view by epyT-R · · Score: 2

    1. No it doesn't. It says that users of binaries with gpl code in them have a right to the source upon request. The vendor has the right to ask a small distribution fee for this.

    2. Well, yes, it is viral. So are many closed source licenses. This virility protects the freedom inherent in any original code remaining in the program after the changes. You would say this to anyone wanting source access to closed applications, right? For those, you charge money, for GPL the cost is your code, which then keeps the application and its evolution free for others to use and modify. The point is to maintain this freedom of access and use to everyone. If you don't want to distribute your changes to a GPL program, don't distribute your binaries. You could also ask the author for an alternative license as he still holds the copyright.

    3. No, it doesn't. You can GPL software and charge money for access.. What you can't do is limit what the user does with it afterwards other than demand he respect the GPL (thus you get access to your user's changes). Obviously this won't work if your goal is to drive value by artificial scarcity of access. If so, that's fine, but then the GPL isn't for you. It's not a danger. Just don't use it. As far as other assets go, the author can choose what parts are licensed in any way he chooses. The GPL does not prevent this, nor can it. For example, the quake3 source code was released under GPL by id software, but the data files were not. Years before this happened, id software distributed quake3 binary only on linux and was in full compliance. There are plenty of binary only applications that run natively and legally on linux/gnu userland.

    4. If so, then the only 'license' that works for you is public domain/no license at all. You're welcome to do that with your own code.

    The rest of your statement is based on your broken presuppositions. It also sounds like you're demanding that OSS developers release under a BSD like license just so you can take their code, use it to compete against them, and give nothing back. Well, again, it's up to the authors to decide, but obviously a bunch of them want to be paid for their work in 'code' rather than in cash in order to keep the project's evolution open. If you're an end user who has not distributed changed binaries, you are not under any obligation to anyone.

    Btw, there is also the LGPL, which allows dynamic linking to GPL libraries without having to comply with the GPL for your application's source. This has been around for decades now, so I am surprised you haven't heard about it. Many GPL libraries are licensed under this. The GPL does not demand that code compiled with GPL tools be GPL'd. Where do you get this bullshit misinformation?

    As far as anti-cheat/gold mining goes, closed source binaries don't seem to do much for that either, since 100% of that has been done on closed source games, and 100% of it has been defeated. Blaming the GPL for this is mind numbingly stupid.

    I see nothing wrong with kickstarter projects. I think it's great. The community gets an open product that will last through many generations of hardware/platforms as long as there is interest, and the developers get paid for their work. Sure beats paying over and over and over and over and over and over and over again for SaaS crapola that basically has no value as it could disappear at any time.

  9. Re:Gnu.org = biased view by Minwee · · Score: 2

    "Christmas" has an H in it, Mr Baldrick. And an R. Also an I and an S; also a T, an M, an A, and another S. Oh, and you've missed out the C at the beginning. Congratulations, Mr Baldrick! Something of a triumph, I think -- you must be the first person ever to spell `Christmas' without getting any of the letters right at all.

    Oops... Wrong conversation. Let me try that again.

    Congratulations, Mr. Coward! Something of a triumph, I think -- you must be the first person ever to write a GPL troll without getting any facts right at all.

  10. Re:Partly BS by msobkow · · Score: 4, Informative

    You are absolutely wrong about point 2. A GPL licensed library does not give you the freedom to license your code "however you like." The LGPL does that. Don't confuse the two.

    If you use a GPL library, you're required to use the GPL license for your code as well. This is not an accident or a "mistaken interpretation" of the license. It's clearly stated and has been known since the first version of the GPL license was released.

    --
    I do not fail; I succeed at finding out what does not work.
  11. Re:Gnu.org = biased view by jythie · · Score: 2

    Having worked in the game industry using GPL components, including respecting the spirt of the licenses and giving back to the community, I have to say.... no. There are plenty of ways to use GPL components within a game without having to give out the parts game parts, GPL is fairly explicit about what boundaries the license crosses and which it does not.

    Now, granted, we did not allow GPLv3 based projects to touch our code, and I would argue that GPLv3 can be pretty bad for people who want to integrate it into larger software packages. Great for people who do server stuff since it was built to handle that crowd, but yeah, games and such, not so great.