As the mail-list discussion has raised a number of interesting points, here's a link to a paper I wrote last year which might be of interest;
http://www.payback.ie/jh/Trusted_email_using_Hashcash.pdf "Trusted Email using Hashcash - Email spam is a growing problem; current anti-spam techniques do little to minimise the volumes. Would a new approach such as Hashcash proof-of-work prove a viable and universal standalone solution?"
The research method I used involved spammer breakeven (profitability) calculations mapped against Hashcash stamp bit-size. I've also done some analysis on requirements of an anti-spam solution, and categorised existing approaches to the spam problem. Including a broader look at policy/trust proposals.
This is not a rebuttal to 'proves not to work', but takes a different approach to the analysis.