I would first like to say I've only downloaded the beta last night, but so far I really like what I see. I'm using the tutorial file to experiment before I move on to the more complicated matter of converting my existing files to this format.
That said, I did come across something that could be a bug. I've been tinkering with the metadata lists (brilliant btw). While trying to add the new metadata items to the outliner columns, the last one I tried to add rearranged the entire column set. Some columns hadn't even been selected. I thought I was losing my mind until I noticed the checked columns that I remembered had been selected (see pic)
Scrivenerprob.jpg (75.49 KiB) Viewed 349 times
I only tried adding 3 items as columns: 2 lists and 1 custom date. The custom date column was the last selected and was when the column set changed.
Minor suggestions: So I've noticed with my use of (old) Scrivener that I use the inspector as a glance ref. I don't want to spend too much time parsing the info. With metadata, there is a line that separates created/modified dates from "include in compile" and section type. Personally, these bits of info are already separated by the difference in input. However, it takes a second or two extra to parse the page title and "general metadata" tab from the info listed above. Perhaps move the line to separate the tab from the information. Obviously, this is how I use Scrivener and may not work with your vision but I figured I'd mention it. (my suggestion see pic)
Scrivenersugg.jpg (16.42 KiB) Viewed 349 times
The second suggestion also has to do with metadata, particularly lists (I'm in love with them). When creating a list, there's a textbox for: "None" item title. This seems to be different from the metadata title, but is treated as a list item which is VERY confusing. Does the "item title" serve as the default item? If so, wouldn't it be less confusing to label it "None" as default item instead?
Hope at least some of this was helpful to the beta process.