Many browser crashes can be directly triggered by an attacker from a malicious web page, which makes a reasonable argument for calling such problems a "vulnerability" or some other term that implies some impact on confidentiality, integrity, or availability. However, as described, this particular issue seems to require manual steps on the user's part, and the attacker has little or no role in the "exploit," barring social engineering. There seems to be an implicit agreement that self-inflicted DoSes don't cross any security boundaries and thus are not security relevant. The question, then, becomes whether this particular issue is more of a "bug" than a "vulnerability" ? If it's a "vulnerability," then how do you distinguish this from other problems that are just "bugs" ? - Steve
