We study Rule-based Artificial Intelligence in the legal sector. We wish to see how various legal principles can be warranted in, for example, automated law enforcement. We observe that dialogues with algorithms could be an important step to safeguarding legal principles for legal software. Moreover, we observe that dialogues are often possible once formal methods are used. Dialogues with software cannot guarantee all legal principles. As an example we mention Transparency and Legal Certainty. These principles are then compared to each other in the realm of algorithmic law. Computational hurdles imply a clear separation of these two principles.