Flutterby™! : .Net 2003 launch 2003-04-24 17:03:10.003429+00

.Net 2003 launch

2003-04-24 17:03:10.003429+00 by Dan Lyke 9 comments

Spent the morning like I do most mornings: Waiting for Microsoft. Except this morning it was standing in the rain for the 2003 .Net launch festivities. Waited in line (they had the registration information, you'd think someone would run a histogram on last name distributions and optimize the badge distribution process?) for my badge, then wandered around the vendor areas while everyone was waiting for the big morning session.

Whooo, boy. Microsoft is the COBOL of a new generation. If there's going to be any revolution to rival the PC or Internet eras it's not going to be started in that corner. But at least they'll build business forms like nothing else.

After a while of hanging around that mob, I decided that watching Steve Ballmer ("Dance, monkeyboy, dance!") wasn't going to do it for my day, so came back to the office to do stuff this morning. I'll go back for the "technical" sessions later.

("Technical" in quotes because of my experience at previous such events, notably the Direct 3D one where the alleged tech guy was pitching FUD about how they didn't use OpenGL because conformance required some things to be slower, although at the time their OpenGL implementation was faster than their Direct 3D implementation, and couldn't answer an audience question about why you'd want a normal per vertex.)

Yesterday went to the Embedded Systems Conference. I'm working up a longer report with lots of links from that, it was both less and more than I expected.

The presentation was by Scott Stanfield of Vertigo Software. Do programmers these days really not have to know that a shift is a power of 2 operation? Anyway, on to the bullet points:

Do programmers these days really not have to know that a shift is a power of 2 operation?

If I had to hazard a guess, I'd say that many (most?) programmers these days don't have any idea what bit shifting is. I have a rough idea myself, but I'm not intimately familiar with doing it (never needed to). Today's programmers are taught to avoid situations that would result in a need for bit shifts. (Okay, maybe not all of them, but I've gotten this from an instructor (or a book, can't remember which) within the last month.)

Hmmmm... "Kids these days" and "back when I was a boy, we knew how to use the LEA (Load Effective Address instruction) to do certain constant multiplies faster".

Perhaps, then, it is just that modern programming is largely no longer interesting to me, and I need to go off and find new challenges. Actually, Phil and I were talking about this yesterday, about another coworker who's great at building interfaces and glue code, exactly the problems that all the tools of Visual Studio .NET

try to expedite, and we'll happily leave him that work and do algorithms ourselves, which is largely a task that Visual Studio .NET

just gets in the way of.

good god- are you quite serious, Shawn? I guess I'm more of a dinosaur than I thought. What ARE today's youngun programmers expected to know?

The kids in my LUG are into scripting languages - mainly perl and PHP, with some flavor of SQL for web pages. I'm trying to mentor the ones that are interested into doing more C-level stuff. Some of the brighter kids are doing C/C++ in an AP course.

#Comment made: 2003-04-27 03:47:50.766957+00 by: Shawn [edit history]

Mars; Java, APIs and how to put together pre-packaged building blocks. The basis that was presented for not using/needing bit shifting was the existence of the boolean variable type.

Of course, I may just have a skewed perspective, since I've never actually spent time around hard-core C/C++ coding or instructional environments.

I haven't used VS.NET very much but I mostly used its predecessors for syntax highlighting, invisible compile/build processes and cool auto-indenting. I'd agree with you, Dan, that most of the other stuff (Wizards, etc.) just gets in the way of what I want to do. I code because I love to write code. Most of the people I've wound up around for the last several years are coders because of prestige and $.

I've been helping a fellow math student through Abstract Algebra. He's a double-major in computer science. He said he wanted to become a game programmer, so I gave him a copy of Tiburon/EA's initial programmer test. He seemed rather forelorn. It amazes me to see people in computer science degree programs complain about writing code. And many of them have very little clue of how the high level stuff they are told to regurgitate really functions at the low level. Another math/compsci double was complaining about a program he had to write for numerical methods. The professor gave him pseudo-code!!! It's basically a job of trascribing... Grrr... Can't wait until I get into numerical methods next Fall. Hopefully there'll be some cute girls in the class (yeah, right... the class is all math and compsci majors...)

I'm not going to begin to comment on 2003 .NET. I'm glad I'm far enough out of that loop not to even be concerned about the latest iteration of Microsoft's APIs. But I will say that if you get close enough to Microsoft's developers you can actually effect change in the API. When I was working with DirectShow to play back MPEG2 streams in the multimedia engine I was working on, the DirectShow team was blown away that I was (still) coding in C. The engine was cross-platform with the Mac and MetroWerks tended to barf on C++ code (especially code that Microsoft C/C++ liked). So we had a company-wide no-C++ rule. I had everything working beautifully with the Beta DirectShow 6 API. When Microsoft released the final API, they rearranged the entire API so that I had to create a couple COM objects using their classes to do things that worked perfectly before.