Window Snyder has the somewhat offbeat title of "chief security something-or-other" at Mozilla, where she is responsible for overseeing efforts to boost the security of the company's open-source offerings, including the Firefox browser.

Snyder joined Mozilla in 2006 from Microsoft, where she was a security strategist who worked on the company's security-driven Windows XP Service Pack 2 update and other products. In an interview, Snyder talked about the state of browser security today and how companies must work together.

What is the biggest threat to users today?

My big concern is still the individuals out there trying to lure users into malicious sites, whether they're posing as a bank or a site they're familiar with. These things are painful. Users really feel it. The attacks come in through email; they come in through websites. It's these broad-based attacks.

Do you see phishing attacks escalating this year?

They'll get worse as long as the methods they're using are successful. As a software development community, we can identify ways to mitigate these attacks ... so they'll eventually move to something else. We need software that is strong enough to mitigate threats that attackers haven't thought of yet. It's about building a multi-level security architecture so you're building in defence in depth. You're not building a defence against a security risk. You need multiple levels of security within an application because you don't know what's coming next. It's about building a secure application.

How far away are we from getting that kind of multi-layered secure application?

We currently try to build Firefox that way, but there will always be things that no one has thought of yet. It's impossible to build a perfectly secure browser. That's not the goal. The goal is to build the safest browser we can. It's an ongoing process. It's not a goal where we'll say, "OK, we're done."

What's the biggest security problem with browsers today?

One of the things most difficult about browser security is that all browsers are designed to take content from an unknown site, and some of that content is going to be executable. Whether it's a Java applet or a Java script or ActiveX control, it's executing on a user's browser. You're enabling these robust applications on the Internet, this rich Web experience, while keeping the user and the machine safe. That's a pretty difficult thing for all of us to do. Software exists to support some function that enables a user or business. It's not for security to dictate to what the user should do. That's not how it should work.

IT keeps warning users to be cautious and not open attachments from unknown senders, for example. But users keep on doing it. Are users just dumb, as some experts say?

Users are not dumb. I get upset when I hear that. They're trying to accomplish a task. If the security interface gets in the way, or if the security interface isn't easy to understand, they're going to go around it, whether it's clicking through a certificate warning or trying to follow an email link that says your bank account is going to be frozen if you don't confirm your personal information details on this page. For all these things, the user is trying to accomplish a task. The attacker has created a sense of urgency. The easiest thing in front of them is to click on the link right in front of them.