Adobe has offered more details of the 'sandbox' security feature it plans to implement to secure its hugely popular but often-attacked PDF Reader software.
First announced last July, the latest description put out by Adobe's security development team makes clear that Reader's new 'protected mode' will be no mere bolt-on. This is starting to look like a ground-up re-design of how the program operates, almost from scratch.
This effectively relegates Reader to a new rung of privilege below that if the system user, which stops the application simply accessing key parts of the OS such as the Registry or file system as it likes. Instead all such calls will have to go through a trusted broker process if they want to communicate beyond the sandbox.
The new design won't stop exploits targeting Reader but they will limit what can be done from within its confines. At the moment, that is more or less anything the attacker wants, including being able to take over the system.
"The challenge is to enable sandboxing while keeping user workflows functional without turning off features users depend on," says Adobe's blog.
As the developers admit, the potential hole in security is always the operating system itself, which can still be compromised, although exploiting such vulnerabilities is as easy as it easy a few years back. Microsoft's software development lifecycle (SDL) has tightened up code security.
The first version sandbox will also not protect against read access to the file system (which allows data theft) or registry, or restricting network access, but future versions will look at this aspect of security.
Adding defence mechanism to specific applications other than browsers is an unusual approach to application design, but Reader's security troubles have gone beyond that of most applications.
The company's regular security updates have become almost as big an event as Microsoft's own patch Tuesday.
The timescale for protected mode to appear in Reader has still not been confirmed.