You'll certainly take a resolution hit shooting anamorphic in anything other than the DXC 10 unless you use an anamorphic lens, which is pretty much the proper way to shoot 16:9 anyway. Shooting 16:9 is not just a matter of changing aspect ratio though. You need to be aware that the whole look and feel of a shot changes as well. Panning rates are also affected, in general the wider it is the slower you should move it.

Shooting progressive scan will give you higher resolution, about 29% more compared to interlaced, if you're planning on upscaling to 720p then you're going to be well infront with progressive scan. Not only is the camera getting a higher resolution feed to start with you're not taking another resolution hit de-interlacing during the upscaling.

Sadly the XL1 doesn't shoot progressive scan, as far as I know Vegas does as good a job as most things at de-interlacing. There seems to be better technology available but it's rather expensive.

> Sadly the XL1 doesn't shoot progressive scan, as far as I know

You mean it's faked progressive? It does shoot progressive, but I don't know what method. So do you mean you don't know if there's even a "progressive" setting (which there is), or that it says "progressive" but isn't really?

Keep in mind progressive scan is different than the "frame mode" of cameras such as the XL1. In such cameras, shooting "frame mode" would result in a lower resolution image than if you shot interlaced.

If you are already planning on have a 'widescreen' aspect ratio, then it kinda depends on your camera and lense as to what your options are. For some cameras, you can get an anamorphic adapter which "squeezes" a 16:9 image into the 4:3 ratio of the CCD (given that most DV cams have 4:3 CCDs). Also, some cameras can "sqeeze" a 16:9 image into the 4:3 aspect ration themselves. What's the difference?
1) A true anamorphic adapter captures a 33% wider angle of view than what a normal lense would. Therefore, you actually capture more of the scene on tape.
2) Cameras that simply "sqeeze" the image simply use the top and bottom areas of the CCD that would be unused anyway if you chose to simply crop it off by adding black bars in post. Therefore, the image area will have a higher resolution than the same scene shot full-frame and then cropped in post.

What's the best option? It's really your call. With an anamorphic adapter, you're getting a TRUE 16:9 image and the wider angle of view that comes with it. Also, all the unused areas of the CCD (which would be cropped off with letterboxing anyway) are put to good use. With the in-camera sqeeze, you don't get true 16:9, as your horizontal angle of view is exactly the same as if you shot full-frame. It will just be a higher resolution than if you had shot full-frame and cropped the image in post.

I think Jive answered the question, at the prosummer level only the DVX100 shoots true progressive. Be warned though, if you don't understand progressive you can get into a lot of grief, think of it as shooting at half the frame rate. The combination of 16:9 and progressive means you've got to deal with a lot of issues in how you shoot.
Yes you CAN get stunning results and they'll upscale well to HD but there's a BIG "but" involved.

As for deinterlacing, Vegas handles deinterlacing much like any other program will. This basically means that you lose have your resolution when deinterlacing interlaced footage. There are some programs that more intelligently deinterlace footage (most often involving some sort of motion analysis of the footage). The best one I've had experience with is Re:VisionFX's FieldsKit Deinterlacer (link)