Knowing .NET

I'm very excited to announce that as of today, April 1st 2004, I've become the world's first micro-outsourced developer. I've hired great developers in Bangalore (GMT +4.5), Cameroon (GMT +1), Brazil (GMT -3), and Vanuatu (GMT -7). My job is to write the unit tests. I post them as binary assemblies to our Groove workspace at any time of day or night, "injecting" them into a 24-hour implementation workday. After several low-profile expirements with Rent-A-Coder "open bids," I've determined that this "4/1" configuration provides the best balance of quality, reliability, and cost-effectiveness. I pay piece-rate, but my guys are averaging just over $4 an hour, with which they're delighted. Meanwhile, I'm averaging 1230 lines of clean code per day (at 30 lines of code per function point). At this rate, I am anticipating shipping my first application (a game for the Tablet PC) in time to compete for the $100,000 ISV prize, with out-of-pocket expenses totalling around $8,200.
10:45:12 AM    comment []   trackback []

After downloading the gigantic Whidbey Preview (available at MSDN Subscriber Downloads) and attempting to install it on my sacrificial machine, I reliably get a problem while installing .NET Framework 2.0: "RegSvcs.exe triggers "Memory at 0x00143cb0 tried to write to 0x00143cb0. The memory could not be 'written.'" This happens whether I use the main installer or try to install the framework from its subdirectory. YMMV, of course.

9:31:16 AM    comment []   trackback []

IronPython doesn't (yet?) use lightweight code generation, a forthcoming facility in the Whidbey CLR, but Joel Pobar shows "Hello, World" using LCG.

9:13:14 AM    comment []   trackback []

Jim Hugunin has posted a paper on IronPython, a fast Python implementation for .NET. This has triggered a cascade of posts.

The money quote: "[A]s I carried out my experiments I found the CLR to be a surprisingly good target for dynamic languages...."

More: "High system performance is the end result of hundreds of small decisions rather than a single large one. Much of IronPython's performance comes from careful consideration of performance in each design choice...."

Essentially, instead of "just" getting the semantics of the language right, Hugunin strove to use native CLR facilities whenever appropriate and provided alternate "fast-paths" for common situations (such as for function calls with a fixed number of arguments) while providing for the more general solution with slower code.

While compiler writing is the rocket-science of computer programming (although game programming comes close...) Hugunin's tactics don't seem unreasonably burdensome.

8:43:58 AM    comment []   trackback []

No, not really. That's just April 1st shorthand for my point.

Peter suggests an interactive drawing tutor as a great application for the TabletPC, a more dignified concept than my "Draw Draw Revolution" game. He suggests text and a voice-over, simple animation, some way of evaluating the learner's input, etc...

I was struck by the thought that there's no commodity tutorial-building software. There's Authorware, but Macromedia doesn't even put that on the front page of their Website anymore and it costs $3000. Meanwhile, we live in PowerPoint Nation. The last time I looked at Authorware (admittedly, probably 3 years ago) it seemed little more than PowerPoint with test-building tools. My adventures in educational software led to my conviction that evaluation is a very significant part of education (a pretty amazing conversion for someone who spent untold hours giving and attending professional conferences: when was the last time you took a test at a seminar?). Of course, one can build a tutorial in HTML or Flash or PowerPoint or C# or assembly language, but what I'm getting at is that surely there's a market for software dedicated to tutorial design.

And here's the thing: it has to be designed by great teachers and built by great programmers.

8:36:52 AM    comment []   trackback []