There are a couple of reasons to use intermediate CODECs.
One is simply to speed things up. If you’re shooting to AVCHD or some other AVC/H.264 format, you’ve probably noticed that it takes substantial CPU power to edit. Intermediate CODECs generally file size for CODEC complexity, so they’re faster for editing.
Another is layered processing. If I take an AVC video in, edit it, and render it to my final format, no big deal. Technically, there may be some quality loss, but it’s not something you really worry about. However, if you’re processing in stages for some reason (for example, I’ve done animations that took several hours per minute to render, with just half of the project loaded into Vegas), maybe using external tools for processing as well, most of the intermediate CODECs are less “fragile”; they don’t degrade quicky with re-encoding.
And finally, you might want to use the alpha channel. Highly compressed CODECs can’t really preserve transparency, since the effect of compression on the alpha channel can be highly visible. So your choice for preserving transparency is either RAW or a few of the intermediate CODECs.
As for interlaced video — if you render to a non-interlaced format, you won’t technically have interlaced video anymore. But as you’re starting with interlaced video, you’re still going to see the effect of that original interlacing in your output video, even if the format itself isn’t interlaced. You can eliminate this, to an extent, by setting your de-interlace method in Vegas (project preferences) to blend or interpolate the interlaced fields, rather than simply copying the straight from 50/60i to 25/30p.
-Dave