I fail to understand why pdf is not a purely representational document format (i.e. without scripting and executable content).
Print it out, they scan it in, I can't imagine how much time, money and hassle it saves. Allows for fast refunds as well.
So, given that and other useful form type pages, how could we properly manage allowing Javascript for the masses?
At this point the specification is not going to just remove or change the problem elements simply because outside the context of the Internet, their not bad. The specification merely offers flexibility to those using or creating PDF files. It is the abuse of that flexibility that make it extremely difficult to identify "known bad" files or protect against them.
What I was getting at was a centralized approval system similar to an issuing certificate. In other words, to use certain functionality, you would need to request it from a trusted entity. This of course would be outrageous and likely to be circumvented, but it never hurts to dream.