[bitdiddle]

If one looks at a graph of MS stock from 1986-2000 and compares that to 2000-2009 it shows the stock has basically languished. MS is a cash cow, slowly dying. They were the first and perhaps only company ever able to successfully leverage large network effects and create a proprietary lock-in model that has enabled them to dominate the desktop for a long time. Moreover it's so dominant that the only thing now worse than Office is Google Docs :)

Their business practices, particularly with respect to extortionary illegal contracts with hardware manufacturers were only finally stopped with a government anti-trust suit against them.

Although it may be enticing to spend a few years with a slick point and click IDE such as VisualStudio and pretend you are programming, as anyone who has ever done so knows at some point you run into something broken and it's tough shit because you can't fix it.

But you can relax, you are correct. They were never able to "own" the net and they don't own the pipes (which I've read is the real reason Buffett never invested in MSFT, he's a very long investor).

Please don't think ill of those who lived through those years. About three years ago now I swore off windoze for good and my health has improved considerably. I truly hope others have that same opportunity.

I might be early, I often am, but if you have the stomach for it I'd say MSFT is a good short, even if you like the stock it can be just as profitable on the way down as it was for those on the way up :)

[Tarks]

"Although it may be enticing to spend a few years with a slick point and click IDE such as VisualStudio and pretend you are programming, as anyone who has ever done so knows at some point you run into something broken and it's tough shit because you can't fix it."

Really? Pretend? Really? That's what you think? I hope what you mean is the potential for over-reliance on wizards, which is easily countered by, oh, I don't know, learning what they're doing so you can accomplish it on your own, just 20-100x slower. Then use the wizard.

I'm sorry to respond to flamebait but I use visualStudio and emacs, compile by pressing a button and using make. As long as you understand what's going on it's all good.

I also think it's strange that in the Linux community there seems to be such a negative vibe around anything that even slightly raises the level of abstraction on the "tooling" side, yet on the languages side it's mostly a good thing, in my eyes they're both just tools, as long as understanding is sound then bring on the productivity boosts, again IMHO.

[jerf]

"I'm sorry to respond to flamebait but I use visualStudio and emacs, compile by pressing a button and using make. As long as you understand what's going on it's all good."

While I don't "hate" Microsoft, as a professional programmer who has worked in both Microsoft's world and the open source world, I am very glad to be able to work entirely in the open source ecosystem precisely because there is a limit to how much you can understand. You hit it when you hit the closed source components.

I grant you that having source for everything may not be everyone's cup of tea, but being able to dive into the source of everything has saved me so often that it results in a qualitative difference in the experience of developing. I haven't spent two weeks trying disparately to figure out how to get around some bug since I used a closed-source PDF generation library three years ago, but this was a routine experience in the Microsoft world. (Not always two weeks, granted, but it can be!)

We've fixed the kernel, read the source of glib and other fundamental libraries to figure out why our things weren't working, hacked up other things in ways that are two-line fixes in what would be closed source in the Microsoft world but would be huge masses of unreliable code to work around at a different layer.

(It should be pointed out that this is simply an effect of having source and the right to build it and use the built result, which some permissive commercial licenses do permit, but my understanding is that this remains the exception and not the rule, with most people still working under the idea that opening your source is tantamount to handing out the family jewels.)

[bitdiddle]

Actually I used all of these tools also but what I meant was if you build a tool any idiot can use then rest assured every idiot will use one. It's all good if the goal is leveraging the skills sets of the folks you hire. For programmers who really love to know things all the way down these tools are at best boring, and many hackers avoid them because when you really need them for something and hit a bug you're dead. There's nothing worse than going to upper management and saying I got stuck because VisualStudio won't let me do X.

This wasn't meant as flamebait.

[Tarks]

" if you build a tool any idiot can use then rest assured every idiot will use one"

True, but not making something because idiots would misuse it isn't a good road to go down either.

"For programmers who really love to know things all the way down these tools are at best boring"

Do you mean boring in the sense that they just work and once you understand them they do exactly what you want, or boring as in you don't care it exists and you'd rather write it yourself, boilerplate included?

For me something that helps speed up a 20 minute job into a 20 second job is pretty cool, speed increase is one of the fundamental advantages of using computers.

"There's nothing worse than going to upper management and saying I got stuck because VisualStudio won't let me do X."

I haven't been hit by that, though I can imagine it leaves a mark.

"This wasn't meant as flamebait."

Roger, remark retracted ^_^

But now I'm curious about ya, When I only knew c++ I would sit and smugly think about how what the java people were doing wasn't real programming because they weren't managing their own memory, they were letting some program do that work. I now consider this a completely silly opinion.

I think control/productivity tradeoffs like garbage collectors are a much harder decision than, say, using a code generator and tweaking the output, as you the programmer have a lot more control over the generated code than you do a VM. What say you?

[rbanffy]

"True, but not making something because idiots would misuse it isn't a good road to go down either."

Erm... I am not sure about that one. Idiots cause a lot of harm.

The speed increase you get when using a wizard that makes code you don't understand you pay back with time and headaches when you have to fix the code you didn't write that no longer runs against the current runtime you are using.

As for Java and C++, they both make writing programs hard. The edge Java has is that it makes writing the bugs much harder.

[bitdiddle]

"The edge Java has is that it makes writing the bugs much harder."

well said!

[bitdiddle]

agreed. I mean boring in the sense that I don't care how my car works.

control/productivity tradeoffs like GC are indeed a tougher decision. For a real-time trading system I would think a JVM might be a disaster, though Erlang/OTP also uses GC so it largely depends on the app and where the GC enters into it. Code generation is interesting. Statically typed languages like Java, enable IDEs that can generate a lot of the boilerplate, through annotations, etc.. but there is also the issue of control/productivity and the economics of programming teams that enters into it. I haven't looked closely but I think GWT is a good example of where folks can leverage their Java assets by generating javascript, for which rich toolsets seems to be lacking.

For what it's worth, having been raised on CLOS and Lisp on a Symbolics, C++ was a step backward and I considered working with it for 3 years a net loss. As I understand it C++ was an attempt to codify what folks were already doing in C with macros. As some famous programmer whose name I forget said, "Java took us about half way back to Lisp".

[encoderer]

Slowly Dying? Please. I can't remember which blog these numbers came for, but this "MSFT is Dying" meme has been circulating for a while and there's a post out there with a fantastic treatment of the notion..

The gist is that Microsoft, as it's supposedly "dying," has gross revenue growing by BILLIONS each year. Yes, their stock has underperformed their sector. But their revenue hasn't. And a company can fudge profits 7 ways from sunday. But you can't fudge topline numbers without out and out fraud.

The author put it in an interesting way: Microsoft revenues grow the size of 1 Adobe or 1 Yahoo!, each year.

There's not a single American (nay GLOBAL) company that wouldn't love for a "death" like that.

And about Buffett... he's often written that he doesn't invest in a company he doesn't fully and completely understand.

[gatormax]

Does John Carmack count as a real programmer?