Slashdot Mirror
Would You Date Microsoft?
teslatug writes "Channel9 has an interview with Bill Hilf of the Open Source Software Lab at Microsoft. Hilf argues that the majority of companies advocate open source solely so that they can drive customers to their core business, which is not open source. He calls this his 'donut theory.' Hilf also sees RedHat in this model, with support being their core. He compares this to dating, where you have to offer your date value in order to entice them. In his view, Microsoft offers developers a platform where they can make money selling their software. The virtues of 'free as in freedom' and the value of open source to the desktop users are skirted, but he makes an interesting point about big businesses like IBM and Oracle."
Well folks, i've been meaning to write this down anyway; here seem's like the perfect place.
Now, I am a 100% Win fan. I love it; things just work. But, I have made the switch to Linux (Fedora Core 5) at home, seeing as it does 99% of what I want. After a couple of months of constant, un-interupted use, my biggest issues with Linux are broadly thus:
1. No fecking media support! I get XMMS inform me on first attempt at playing an MP3 that it won't because of licensing conflict. Wtf? Codecs for avi's and DVDs were a simular story; all had to be downloaded via yum (bloody excellent tool!). Seriously; not good, but fixed in the end.
2. Why the hell do I have to install a new kernel? Why? I've never had to on Windows - why is Linux different? Is it so buggy? I installed with a factory version something ending 054. Now I have something ending 122 I believe. I did it ok, but that's not the point I'm making; were there really 68 cock-ups so great in the kernel build from release-time until that now they had to re-release 68 times? I'm guessing probablly not, but still.
3. Point 2 also breaks my nvidia drivers. I don't want to re-compile new drivers everytime there's a new 'patch'. For the love of god, why?!
4. X-Windows. What a mess. Why do I have to tell it my x & y refresh rates for my monitor? Windows just 'knows'. Many more things here I feel that X-Windows should just 'know' - the number of buttons on my USB mouse for-instance. If Windows can do it, there's no reason why Linux can't. Also, X-Windows 'feels' slower than Windows. I'm sure there's good reasons for this, but I don't care; Windows is snappier.
5. Lack of decent file-browser. The best I've come across is Nautilus in a mode that resembles Windows Explorer. It'll do for now, but as far as I'm aware, offers no context-sensitive menus for applications (like the Winamp "Play in Winamp" right-click menu on folders.
Actually, I think that's largely it. In all, Linux has, and is continuing to be great fun to play with. So many cool tools - yum being one of them. I'll stick to Linux @ home; it can only get better, but I'd be interested to know what people think of the above points - any suggestions maybe? I want this to work after all...
The virtues of 'free as in freedom' and the value of open source to the desktop users
what virtues? He expects the article to touch upon these points, but to many people they have not been sufficiently justified.
I've been using open source software for years, and have heard many people talk as if there was some moral imperative to release software under the GPL, or other oss license. Catch phrases like "free as in freedom," and "information wants to be free" are bandied about, and it is generally implied that commercial software developers are evil in some unspecified manner. However, these attitudes have never been justified to me with anything more than rhetoric and metaphor.
Slashdotters, maybe I am a fool. It might be that the moral imperative behind open source is only so obvious that no one can be bothered to write it down. However, I beg your patience and ask that someone take the time to explain it to me.
Now, to be clear I am not asking how open source helps to develop high quality software. I am already convinced on this point. I am asking for a justification of the commonly observed attitude on slashdot that open source developers are "good" and closed source developers are "bad" in the moral sense. I am asking for a justification of Richard Stallman's position that, as I understand it, there is a moral imperative to develop software under the GPL (or similar license).
Furthermore, as some suspect that I am already clearly quite daft, let's avoid using metaphorical terms or similes in the argument, as they might confuse me. Instead let us use only actual terms. By this I mean that I ask that responders do not derive some moral truth about computer software design by comparing it to plumbing, or cars, or politics (all of which are popular patterns of argument on slashdot). In these forms of arguments we are expected to accept some truth about an unrelated subject as a premise (i.e. you shouldn't send someone to jail for speeding) and from this premise come to accept some truth about computer software that holds a somewhat similar form (i.e. you shouldn't send someone to jail for hacking into their computer). In my ignorance, I often fail to see how the one proposition follows from the other. Often I even imagine that I see semantic distinctions that render the similitude meaningless with respect to the subject at hand. To avoid wandering into these failings in my comprehension, I ask that responders simply tell me why something is directly, without comparison to other truths.
Have at it.
Microsoft's real problem is the trap. They are trapped in a way that regardless what they do in the field of open source everyone believes it was pure propaganda. Which may well be the case.
.Net is a nice consolidation of the former plattform but... oh well... that is not exciting. "The better Java" so to speak.
Today more open source runs on windows than on Linux machines. Content Management means open source. Cluster computing means Linux. Webserver means Apache.
In some areas Open Source provides real advantage. Unlike its competitors Microsoft cannot run a real open source strategy. They cannot use open source for their own advantage.
And what is worse: Microsoft's policy making, its advocacy against open source, against interoperability, money for politicians, money for software patent lobbying and other dirty business provides them with nasty press coverage and they lost the support of the software elites.
What professional developer likes a company which fights for DMCA style laws and software patenting? Microsoft lost the support of developers. Its technology and progress does not excite us anymore. (Oh, I like MDX but that's very old.)
Open source values developers. Developers run open source. No marketing braggarts blur the field. That is why we love it.
You have to install new kernels for the same reasons that you have to buy new versions of windows. (You're not still running Windows 3.1, are you?) You get new drivers, methods, and all those fun things you expect from your operating system. I can write you scripts to mostly automate the process of building new kernels, which should take 94% of the pain out of the process, but it will involve answering stupid questions about new drivers. It doesn't know. Hotplugging is our weakness right now.
I haven't built a kernel from source in almost three years. Most desktop oriented distros provide "kernel header packages" which are basically #include files that match your running kernel. From time to time, I have built third party drivers from source. If a third party driver will build against the "kernel headers", you can build and install it without rebooting; most times it is just the "./configure, make, checkinstall" routine. I used to regularly build the nVidia drivers this way but Ubuntu is good about providing "restricted driver" packages that match their supplied end-user kernels (which are pretty much built in "kitchen sink mode" so you don't have to rebuild to get some obscure option). But even if they didn't, automating nVidia driver build-and-install wouldn't be too hard.
VMware Player is another third party item that works just fine with kernel-header packages. Come to think of it, the only thing I've seen lately that won't build without patching kernel source and forcing a kernel rebuild on you are new versions of the sky2 driver. Even there I managed to get things working without resorting to a full kernel rebuild.
Rebuilding kernels is something I used to futz with a lot. It just isn't as necessary these days, especially if your distro pays good attention to end-user issues.
I used to write software using Microsoft tools. The tools were expensive, and sometimes buggy. And when I encountered bugs in visual foxpro, I couldn't fix them. They were usually fixed in the next version, which had a new set of bugs. It's not terribly buggy, but sometimes one bug can really cause problems.
I also did some VB stuff. They went through three different, slightly incompatible database access classes during my use of it. All were written by committee.
Note that these were the cheap tools, too, I wasn't using sql server or such.
The world of Free software is completely different. I have control. I cannot stress this enough: I HAVE CONTROL. It's considered a myth that anyone can fix bugs, but I have more than once. I remember well fixing a bug in the pop server that I'm using. It would have taken Microsoft or a company such as that a month or more to fix a bug like that. It took me 30 minutes from never having looked at the code to having the bug fixed, patch sent to maintainer.
Now, for the stuff that I do nowadays, not only is the control factor large, so is the cost factor. They are correct that Microsoft provides a platform where you can make money. But that means you have to give Microsoft some of your money. If free Free software didn't exist, that would make sense. However, in the presence of an equal or better alternative that costs no money, it makes absolutely no sense to give Microsoft money for their often inferior offerings.
I have a particular client that I took from another company. It was an ecommerce site, nothing special, frankly. The other company had already billed the client $40,000 and the product wasn't yet working. The client brought me in to help the other bozos with some html. Yes, you read that correctly. So I asked the client for their data set, and three days later showed them a prototype that was more functional than what the other guys had spent three months and $40,000 to accomplish.
I then made them an offer. They hadn't paid for the Microsoft licenses yet, which were going to run about $15K. I told them that I could deliver the entire thing for less than the up-front cost of the Microsoft licenses. In other words, they could abandon everything that the other guys had charged them so much for and still save money. They decided to play both sides, and a month later I delivered the completed site, under budget. The other guys charged them another $40,000 for time they had spent since the last bill, but still no completed site. I don't know if they paid it.
I have found that most companies like those do not inform their clients ahead of time that there are going to be Microsoft license fees to pay. They rather find out afterward. In this case, when the guys found out what I was doing, they went to the client and told them falsely that they didn't have to pay for those licenses, that they could just use a free test license.
There's a lesson there, though. For most larger projects, those license fees are laid out up front (although they are usually dishonest about the ongoing costs, I've found). But think about it. If a client is going to spend $50K on a project, my choice as the vendor is either $50K in my pocket or $40K in my pocket and $10K in Microsoft's pocket. Again, for what? Better yet, I can "undercut" at $45K, still make more money than the other guy and save my client money.
Note to other vendors: keep pushing Microsoft crap at people. I love it when you do. Seriously.
Do you have ESP?
In addition to trying to make open-source business models seem just like commercial ones, as in "they just change the core of their doughnut" (from intellectual property to support services), this Hilf fellow isn't very accurate (honest?) about the actual core of Microsoft's doughnut.
I couldn't play the video (I'm using Linux, so probably haven't got the right codec for Windows Media Video, or whatever it is), but as an economics student, from what you've written, it sounds like quite a reasonable argument. Microsoft's software, or IBM's services, are but a small part of the overall set of tools used to build IT infrastructures. Both the Microsoft and IBM models involve focussing on the particular tools where the firm's management believe the firm have a competitive advantage (e.g. closed-source software IP in Microsoft's case, and principally customisation/support services in IBM's case), and either giving away the other tools, selling them cheaply or allowing partner firms to provide them. In that sense, then, the models are essentially identical, so I'm not really sure what your objection is.
Microsoft's core asset isn't Windows and Office. Microsoft's core asset is their monopoly, without which their whole model collapses (or, if you like his metaphor, their doughnut crumbles).
In economic terms, neither Windows nor Office is a monopoly. The Windows desktop OS (but not Windows Server or Office) has been ruled to be a 'political' monopoly by the government of the USA, but that doesn't mean it actually is a monopoly in economic terms, which it isn't, and never has been. Operating systems are arguably a 'network economy', which means that as the installed base for a given product grows, the value provided by that product increases, but being the leading player in a network economy isn't the same thing as being a monopoly. Moreover, whilst network effects undoubtedly help Microsoft, they only came about as a result of having previously won in competitive markets. Microsoft managed to overcome Lotus in the applications market and the Unix vendors in the server market, despite both of them having initially had substantial network effects on their sides.