Let me start, and then I'll throw it over to Cliff.
Just to give members a sense of this, we've been working since December, anticipating that the committee may decide to extend the number of weeks of benefits. Given the nature of the COBOL-based system we have, it will take a full four months for us to be able to implement an extension in the number of weeks. For something like the sickness benefit that Mr. Blaikie mentioned, to extend the maximum number of weeks is actually a major system change. It's not like the one we're doing right now. It changes the fundamental system logic. A change like that takes a minimum of nine months to do.
The challenge is that with a COBOL-based system, simultaneously processing major changes massively increases the risk that we cause irreparable damage to the system. Part of the issue is that were Parliament to make a decision around a particular change, that would preclude us from being able to do other changes. For example, if there was a decision to extend the sickness benefit beyond the current maximum number of weeks, that would preclude us from being able to do other changes to the system for September. It's a function of the limitations of a system whose code is almost as old as me.