I agree with the points my colleagues made, so I won't repeat them.
However, I will say we definitely have a challenge as far as data publication and official languages are concerned.
Internally, the datasets contain text that our researchers and scientists use and sometimes that text is, by default, accessible in only one language. So publishing a dataset always poses that challenge. There is a requirement to release in both official languages metadata that, under normal circumstances, is available internally in just one language. They may not be automatically recorded in the same way, so that has to be done. But I think the key lies in avoiding the duplication of work and opting for an open approach right from the get-go. That means taking the steps that need to be taken and putting the necessary processes in place to do the work the right way and make the data accessible to Canadians in the appropriate format.