.NET and the Cult of Microsoft

[elfs]

I've been looking for a decent book on doing ActionScript development without the Flash environment, since I have Swfmill and MTASC. But I've also been looking for pointers to the C# development environment for Unix, so while I was at Half-Price Books I picked up a used copy of ASP.NET in a Nutshell, since I tend to trust O'Reilly (although that trust is sometimes not warranted; it took O'Reilly forever to get their heads out of the butts with respect to Python because they were too Perl oriented once; and their XML and UML books have just been awful).

There are some moments in this book that reinforce for me the whole Cult of Microsoft meme. Seriously. There's a chapter in the opening where the author absolutely crows about how ASP.NET is "completely object oriented." In 1996, NeXT, Inc. released the very first web application server, WebObjects, and it was written in Objective-C, a C-based language with Smalltalk-based object orientation that was arguably more "object oriented" than C++ has ever been. Almost every web-based framework since then has been OO; For MS to act as if in 1992 it brought manna to the masses by "introducing an objected oriented framework for web development" is simply ridiculous in the extreme.

The whole section on authentication and authorization pooh-poohs the basics of HTTP authentication and seeks to "route around" it by providing extra, Microsoft-only solutions such as Passport. There's no mention of other, more popular authentication schemes. It's as if, once you drank the Koolaid, nothing else in the world exists or had ever existed: not Apple's WebObjects, not Zope, not PHP, not Rails.

It is stuff like this that bothers me most about Microsoft: the whole "it wasn't invented by us therefore it's not worthy of your attention" attitude that is so all-encompassing as to be a distortion of reality. It's no wonder that C-Sharp, despite its obvious superiority to Java, makes developers itch.

Comments

[duskwuff.livejournal.com]

To be fair, HTTP authentication is generally to be avoided in web applications. It's much less flexible than cookie-based methods, significantly less secure unless you use digest mode (which has only become widely supported recently), and is generally a pain to use.

On the other hand, suggesting Passport as a solution is less than helpful.

[fallenpegasus]

I heard a while ago the statement that "Java is the COBOL of the 21stC", and the moment I heard it, I knew it was true.

Which then caused me to ask myself, "So what is the PL/I to Java's COBOL?". And after a little bit of thought, I knew the answer.

.NET

[kyriani.livejournal.com]

This mentality is very sad to those of us who are doing C# and .Net for a living that could just as easily do Java or Python or Ruby on Rails. The client chooses the .Net platform and we run with it. The people I want to hire for my team are exactly the people who are NOT those who drank the koolaid, though they are depressingly hard to find. :(
*sigh*
The whole C# open source movement gives me hope though... NAnt and NCover are extremely useful in our development environment. I just wish more people thought more about what they said before they say it, there really is a "Cult of Microsoft" sadly.