One factor that gets missed is the quantity of stuff which only ever gets one release. Not just books and music, though they're what I notice. There's an argument that digital copies, like grey imports, reduce the potential profit from sales in other countries, but the international nature of the media companies rather blows a hole in that argument.

But how many new works will still be in print this time next year? The music industry claims that most artists aren't profitable for the company, so what advantage to them is fifty or seventy or a hundred years of copyright?

How many books, first published ten or twenty years ago, can still be a commercial success today?

Yes, let the creator of a work have the big money, if he strikes lucky. But too many media companies buy the copyright, and never let go of something that they don't even try to exploit.

I seem to recall that someone (Lawrence Lessig?) pointed
out that of the books published in the USA in 1932 which
would be coming out of copyright protection this year -- if
not for the 1998 extension to 90 years' post-mortem --
only about 5% were still in print.

There's a message in this, somewhere.

Something else that needs to be taken into consideration is the extent to which commercial pricing of much software is based around corporate buying: that is, Adobe expects companies to buy Photoshop, and treat the purchase as a deductable expense. So: how does someone gain sufficient experience to work with Photoshop in an office environment? Essentially, by piracy. So there's something of a blind eye turned to low-level individual piracy for software that's essential in the workplace, simply because it's the only way to sustain a flow of sufficiently-trained (or indoctrinated) workers. After all, corporations can't usually get away with piracy, given the role of FAST and other organisations in auditing offices for licences. Think of it as an equivalent of the drug dealers' tactic of 'you get the first one for free, and then you're hooked.'