According to Rich Salz’s reading of the WSDL 2.0 spec (Rich makes a living selling hardware to make XML smoke, so I think he knows), the following is legal WSDL 2.0:

#include "wsdl.h"
extern void hello_world(const char* text);

It’s hisjudgmentt that this is not a step forward and I’m having a hard time finding fault with his argument. : )

I’ve been wanting an easy way to capture screen demos and audio for a while and after stumbling across Jon Udell’s post on Movies of Software, I thought I’d give Windows Media Encoder a try. Here’s what I learned:

• I really like giving short little vignettes imagining an audience; I didn’t think I would
• The preparation for recording a screen demo is the same as a real demo, i.e. I want to make sure I know what I’m going to demo before I do it
• The preparation does not include lighting or another person to run a camera or a microphone, which means that I can do it on a whim at 5am on a Saturday morning
• My built-in mic at normal speaking volume works great
• Capturing a particular screen doesn’t work very well because I’m likely to move between windows without knowing that I will ahead of time. Instead, I should capture a region of the screen, ideally something not too large so that it fits well in a video window
• Medium quality may not be good enough. I did 13 minutes at medium quality, which resulted in a 3MB file and good audio, but the screen itself was nearly unreadable. Maybe the screen itself was too big?
• Stop other apps before capturing a screen, otherwise you’re bound to get slowdowns in the capture, particularly in the audio
• Talking in my kitchen at normal volume for 13 minutes will waken my wife (sorry, sweetie)

So, while I wasn’t happy with the results, I’m very jazzed about the medium. Expect to see more.

Jeff Barr of Amazon has just informed me that Amazon now provides a distributed, shared, always on message queuing service which allows you to create and delete queues, enqueue, read and dequeue messages that can stay put for as long as 30 days. It's free right now and I don't quite know how it will be used for porn and spam, but I have every confidence that you guys will figure it out. : )

I've never really seen anyone use speech recognition for real and after watching Jon Udell use Dragon Naturally Speaking 8.0, I'm flabbergasted. I had no idea working with it could be so natural. Wow.

When I first read about MaxiVista, the thing that really got me interested was that it gave me two things that I really want from my Tablet PC. The 1st is to be able to use a slate-only Tablet as a 2nd monitor for my laptop so that I can continue my addiction to keeping my whole life on a laptop, but still be cool like the other kids (I’ve so far avoided multi-monitor setups because I love my laptop too much and I was afraid of the conflicting addictions).

The 2nd thing I get is a result of the first, i.e. if I’m going to use the Tablet as an adjunct to my laptop, then I shouldn’t worry about a generation sometime in the future where Tablet convertibles can replace my laptop, but just get the best slate I can find and enjoy it as a remote viewer/editor of info via file sharing and terminal services.

If you have a device you’d like to turn into a 2nd monitor and you don’t yet have MaxiVista, Scott Hanselman’s posted a 30% off MaxiVista coupon.

Like everyone else in the blogespher, I installed Google Desktop Search last week. And, after using it a few times, I've uninstalled it. Jim’s right; X1 kicks its butt.

I haven't even seen this yet and I want to recommend it. I'm just a big monad and Jeff Snover fan.

Today Ralph Loizzo asked me what tool I use to outline my objects, to which I replied:

The tool I use most often to design my systems is a text editor. I write my client code first against the objects I wish I had to obtain the first order of functionality I’m after, then implement those classes and refactor ’til I’ve reached a state of happiness between the clients and the object model.

In this age of development methodology du jour, I assume my methodology already has a name, but today I dubbed it CDD (Client-Driven Development). Also, I don’t distinguish between client UI code and lower-level code, i.e. just as often as I write lower-level code against an imaginary OM, I design my UIs first, then implement an OM to enable the UI.

Over the years, I’ve become less and less of a fan of the idea of componentizing the functionality you think the higher levels are going to need, as it often results in over-engineering. Instead, I prefer just-in-time engineering, refactoring my design as I need new or different functionality, whether in the client during the initial development stage or during the maintenance stage.

Test Driven Development is very similar to Client-Driven Development:

• TDD: writing the tests first and writing your code ’til the test pass
• CDD: writing the client first and writing your code ’til the client has the functionality it needs

It may even be that CDD is a degenerate case of TDD where the “test” is “the client seems to work.”