The Third Bit

The second revised lecture on Python for Software Carpentry is now on the web. As always, comments and corrections are welcome.

Software Carpentry

According to MyHeritage.com‘s facial recognition software, the three celebrities I resemble most are the director George Cukor, the actor Hugo Weaving (OK, I can live with that), and the dancer Gene Kelly. Numbers 4, 5, and 6 are Kevin Mitnick, Noam Chomsky, and Michael Caine; make of it what you will.

Coincidentally, I’m reading Bruce Sterling’s new manifesto, Shaping Things. In it, he describes a near future full of “spimes”: physical objects with electronic identities, sensors, and communications gear that will turn us all into ad hoc designers. “You won’t look for your shoes in the morning, you’ll google them.” MyHeritage’s site is just one more example of what next-generation search interfaces will look like (and yes, I do believe the rumors that Google’s working on a system that will search for tunes based on you humming or whistling a few bars…).

So, want to be rich, famous, and popular? Figure out (a) what kinds of rich search interfaces will be most useful to programmers, and (b) what kind of programming system(s) will give Jon Udell’s Aunt Tillie the abilitiy to use all this stuff. I suspect the answers will be very similar…

Uncategorized

Adam Goucher has blogged a piece on how to set up hardware for a testing lab. Now you don’t have to reinvent these wheels…

Uncategorized

In last year’s piece on extensible programming systems, I said that the biggest obstacle higher-level programming tools face is the lack of decent debuggers for them. Anyone who has wrestled with a broken thousand-line Makefile, or a complex Apache or Tomcat configuration, knows the problem: you write at this level, but when it doesn’t behave the way you want it to, you have to reason backward from that level to figure out what to change.

This posting in ScottGu’s blog pulls together a few visual debugging aids developed for .NET. Taken together, they show how much more we could be doing to make developers (who spend more time debugging than writing new code) more productive. What we really need, however, is languages and platforms designed with extensible debugging in mind. Ryder et al’s paper The Impact of Software Engineering Research on Modern Programming Language Design (ACM Trans. on Software Engineering and Methodology, Oct 2005) shows that language designers will pay attention to developers’ needs, provided those needs are organized, categorized, given some theoretical underpinnings, and clearly explained. Andreas Zeller’s book Why Programs Fail is a start; here’s hoping others will pick up the torch.

Research

Enter the eBay Challenge today! Win stuff! Become rich, famous, and popular (well, one out of three)! Watch from the front row as post-modern programming turns into the web version of the Gong Show!

Uncategorized

The first Software Carpentry lecture on Python has been revised, and is up on the web. Comments and criticisms are very welcome.

Software Carpentry

A new version of DrProject went live on Pyre this afternoon. If nothing serious shows up in the next 24 hours, we’ll cut the main installation over. It’s leaner, cleaner, tastes better, and has fewer calories—I’m very excited, and very pleased with how quickly Chris, Jason, Sean, and our ever-helpful team of volunteers have pulled it together. If the next two weeks go as well as this one has, we’ll have a 1.0 release in February, and will open up the repository for general contributions then.

DrProject

The revised lecture on Make is now up for comments and feedback. Still no diagrams, I’m afraid…

Software Carpentry

Via Mike Gunderloy, filed under “surreal but true”: a web-based system that will call your cell phone when your laundry is done.

Uncategorized

I have revised the second shell lecture and the lecture on version control for the Software Carpentry course, and added the necessary glossary definitions. As always, comments are greatly appreciated.

Software Carpentry