Because in our cloud system we’re building for all the Executive Yuan and all the administration, we have a flag in which we can just switch and make this discussion board public, right? So a lot of these are we want to feed into machine learning. We want to feed into the ways that analyze these materials and so on. So if you mix the secret information with unclassified information and then produce some aggregate results, then it is very questionable whether the aggregate result is classified or not. According to our current interpretation of the law, any system that contains input of classified information has by definition of output as classified information, which will jeopardize the whole effort, right? Which is why we only work with unclassified open data.