For example, I think drag and drop demos and controls serve several important goals. As I’m sure you all remember there was a time that in order to use a new feature you needed to write hundreds of lines of code. For example I recently had to do some video driver work with DirectShow and to do this one simple thing I had to jump through about five different API groups and it required about 300 lines of COM code. I would love to see a simple one line way of doing this (and frankly, it should be doable in a line).

The drag and drop demo is about showing that we’ve introduced this functionality and its easy to get at. Now this leads to Alan’s other gripe about the code being bad. In some cases it might be, but I will argue that almost all code is pretty bad, regardless of how it was generated! I think it’s easy to point to drag and drop being the problem, but I think if the code wasn’t drag and dropped I suspect that you’d likely find that the code would be no better written from scratch in vi.

Lastly, I found it somewhat odd that .NET developers are the ones complaining about the abstraction presented by drag and drop. To me that’s like getting a diet coke after ordering two Big Macs, large fries, large chocolate shake, and an apple pie. The issue is way larger than the coke.

As someone who has spent a good chunk of his life writing x86, 68k, PPC, and x64 assembly code the drag and drop adds so little to the abstraction space. The fact that most developers can’t tell you the cost of a foreach loop or the cost of a stackwalk scares me more than the use of drag and drop.

And lastly I would end that I’ve seen some solid LOB apps written drag and drop. They were written in a week and in many cases do their job for several years, until they need to be dumped or largely rewritten to deal with drastically changed requirements. To me it seems hard to justify throwing away this productivity because some bad developers write bad code more quickly with these tools.

It’s not hard to illustrate the horrors that can result from uninformed drag & drop. However, I love that drag & drop (and RAD development in general) can help people get started and be productive. All of us learn along the way and drag & drop can help reduce the initial learning curve.

Sometimes I think what Alan is ultimately proposing is akin to teaching a 1-year old to walk by tossing them on a treadmill running at 10mph.

This being said, I understand the hunger from seasoned developers for demos with a bit more meat.

Thanks for having us on the show. Andrew is a great guy even though he’s wrong most of the time. ;-)

++Alan

And I agree with Scott…what a great summary. I just hope the debate lives up to the description. :-)