ARGUMENT 2:
It’s not quite as simple as ‘hence the multiplier.’ The multiplier tells you the relative sizes of the sensors (and thus their relative fields of view) but doesn’t tell you anything about the number of pixels in the sensor. The relative sizes (in pixels) of objects photographed under our lock-the-tripod-swap-the-lens test conditions depend on both.
A 30d sensor has ~8 million pixels. A rectangle the same size as the 30d sensor, if it were placed over a 5d sensor, would only cover ~5 million pixels. If you view them both at 100% on a monitor, the rectangle photographed on the 30d will be bigger than the one photographed on the 5d.
You can crop as much as you like. You can shoot with a portrait lens and crop down to the field of view of a spotting scope. There just won’t be much detail left.
If you crop a 5d to match the field of view of a 1.6 crop factor camera, you will have a 5 megapixel image. IIRC, the highest resolution 1.6 crop Canon is currently 10MP, so the equivalent cropped area of a full-frame, 12MP 5d will have noticeably less detail than the equivalent image from a 400d.
If the rumors are correct and Canon releases a 22MP, full frame DSLR, it will have better detail (even cropped) than an 8MP 1.6 factor camera, but just barely. It will still have less absolute resolution (after cropping, for a given lens) than a 400d. If you’re trying to photograph something that only fills a fraction of the frame with the longest lens you have, and you absolutely want the most detail possible when you crop the rest of the image out, you don’t want a 5D. That’s not what it’s designed for. (You really want a D2X – that is what it’s designed for – but there’s the issue of lens investment, etc…)