What's the Business Case for Microsoft and Open Source?

Friend of perl developers everywhere, Jeremy Zawodny, has an intriguing question: "If you had to explain to Microsoft why they should change their attitude toward Open Source, what would you say?" For more about this, read on... From Jeremy: "If you had to explain to Microsoft why they should change their attitude toward Open Source, what would you say? More to the point, how can Microsoft benefit from better supporting or even adopting Open Source in their business? (Replace IIS with Apache, for example.) Does it make sense for them? Are there ways that they can use Open Source as a competitive advantage without pissing off the Open Source community in the process? Which of their products would make sense on Open Source platforms? How can the Open Source community help Microsoft? Or is this a lost cause? IBM has made it work. Can Microsoft?

I ask these questions because I may have the chance to talk with folks at Microsoft about Open Source. And it only makes sense that I look to the community for input. So let's hear it. Flames won't help. Thoughtful answers and ideas very well could."

Comments

Eh?

[MisterBlister]

Who cares if 'IBM made it work'? IBM is in a fundamentally different market than Microsoft. IBM may as well be in the fish packaging market for all the differences there are between service & support of large iron and end-user software for the home and office.

The sad truth is Microsoft has nothing to gain monetarily from moving to Open Source, and since they are a corporation, money is all that really matters. I don't mean this in a bad way, its just the way it is. There's so many hurdles that Microsoft would have to overcome to make things OSS.

Consider how much their legal dept would get the sweats over a shareholder law suit if they OSS everything and the stock drops because nobody buys software anymore -- they just download the OSS Microsoft code and compile it!

And that's only ONE of thousands of problems. They also have the standard problem of using a lot of code licensed from other people, how do they deal with that? Even if they wanted to OSS their software because there was a good reason, it would cost the millions if not billions in legal fees and programmer time just to get rid of all the licensed code depedencies in their software!

In short, forget about it. Use your energy on something else.

Windows - the Pinto of the 21st Century

[dfn5]

Slashdot has the answer to your question right here. I honestly don't see how one can change the views of Microsoft when they are making claims like that.

Re:Don't Fool Yourself

[timothy_m_smith]

In terms of getting MS-type developers to write Open Source code, you will have target your audiences selectively. I am in consulting and I focus on custom developed software. The kind of big corporate clients I work for, would never let any part of their custom business apps be published as open source. These companies are looking for any advantage over compentitors and so they will not give any part of their code away. For software that is fairly common among competitors, these guys generally will buy some kind of package. So, I think you're going to have to find another group of MS developers besides the ones working at big corporations.

Re:I'm confused

[smagruder]

I did my research just fine. Crystal Enterprise doesn't support Apache on Windows.

Re:OS license cost

[Zeinfeld]

You use VB for RAD, you use C when you need power, doesn't mean one is better than the other, they have different purposes and solve different problems.

Once this was the case, however the real importance of C# is that it finaly merges the Basic and C code development lines. At this point C# provides the full power of C (with some bizare omissions like structure initializers) plus the convenience of VB. The market perception that C# is about Java is only really 40% right, C# is pitched against Java because Java is the language to beat, but the real target audience is Visual Basic programmers who would like a programming language that is as easy to prototype in as a scripting language while still allowing very large projects to be supported.

Equally the suggestion in the intro to the article that Microsoft should switch from IIS to Apache is amazingly clueless. Apache is a great Web server for UNIX boxes, but IIS is a better Web server for NT. IIS is integrated into the O/S at a very fundamental level so that for example the Web server can use the system level file protections to control access to Web resources.

The features that have caused security problems with both Apache and IIS are active code. In the case of IIS three scripting langauages are integrated into the Web server (and more can be added). In the case of Apache the security weaknesses inherent in the CGI design (particularly when a CGI module is written in csh) leads to predictable problems. I don't see that a real difference can be made between the OSS and Microsoft approach here, both groups adopted what is an intrinsically insecure architecture for reasons of expediency and ignorance. Once the feature was in there was no way for the grown ups to take it out again because people used the feature.

I recently started using Visual C#, its the best program development environment I have seen since the VAX LSE. The editor does have some iritating features (like the lack of mouse-less editing), but it does have a lot of cool features like bringing up the template for a method as you enter it - even for user defined methods. The IDE looks and feels like a professional tool, there are few traces of ego-centric features that looked cool to the designer but are not so great for the user - although as with XMLSpy the editor makes the bizare assumption that my preference for editing XML Schemas is through some bizare graphical language of the authors invention rather than as XML schema.

Compare Visual Studio with the UNIX - Emacs - Make IDE and I am afraid the comparison is not favorable to open source.

I am much less interested in open source than I am in extensibility. Unless you want to do a security audit the only reason to want source is to maintain or extend a program. I much prefer a well written and supported extension mechanism than someone chucking a few meg of code at me. The .NET extension mechanisms allow me to write my own language and then use Visual Studio as my IDE for it - and get all the debugging, assistant etc. features for free. That seems somewhat better to me than creating a fork of the emacs and gcc tree for my new language and recreating all those features.

YMMV, but those people who believe that OSS is the one true faith are wrong. There is plenty of room for both models in the market of ideas.

Re:Don't Fool Yourself

[Zeinfeld]

Anyone who says *opensource* is doomed to fail, then say "oh yea, obviously, that whole HTML thing was just a fluke, sure glad it died out in the earily 90's after every READ EVERYONE ELSE'S CODE, LEARNED HTML AND MADE THEIR OWN PAGES!"

HTML was not open source. It is an open specification. There are open source browsers writen to that specification but the specification is not the code.

Also the libwww code written at CERN is not open source, it is public domain. There is a big difference. If you modify libwww there are no limits on what you do with it, you can make the modification closed source.

We did not write the license that way because we were ignorant of RMS's politics, far from it. The license was written that way so that companies such as Spyglass and Microsoft could build systems built on our code if they wanted to.

As for reading people's HTML to write code - thats a bug not a feature. The original idea was that the browsers would have the ability to edit Web pages so the end users did not have to learn HTML. Also HTML was originally much cleaner than the current spec which has countless enhancements added in by Netscape in an attempt to make the spec proprietary during the pre-Microsoft browser wars against Spyglass and the Web Consortium. Thats why we have six incompatible mechanisms to change fonts but none of the standard browsers support math markup - bit of a lose for a technology meant to be for scientific publication eh?