(warning, this sounds and probably is more than a bit douchey, and I apologize in advance.)
The major issue I have with "capriciously applied" rules, (Like "Magic is not science!", when if there is a force in the world, and it can have demonstrably repeatable results from a repeated action that involves it (such as an incantation)--- then quite clearly, magic *IS* a science, resulting in a paradoxical, or nonsensical statement from the DM, etc.) is that the goal of a role playing game, is to ASSUME A ROLE, and the whole point of creating a fantasy world, IS TO INTERACT WITH THAT WORLD.
When a DM simply does not want to permit experimentation or critical reasoning on the fantasy world, it denies the player a true understanding of what they are actually doing, and denies them the ability to actually play a role.
Example: A wizard.
A wizard is *NOT* a sorcerer. (A sorcerer's magic is wild, and a product of their imagination/charisma interacting with their magical abilities.) A wizard learns magic as if it were a science. 'This principle of magical energy interacts with that feature of mundane matter, producing THIS effect'--- etc. This is necessary for a standard spellbook to even exist.
Wizards learn these interactions, and their 'research', is finding more detailed understandings of those interactions, or finding exceptions to the rules of those interactions, in a disciplined and repeatable fashion. It would be in a wizard's nature to want to push the limits of their knowledge, or to push the limits of what is (currently) possible with their knowledge of magic and how it interacts with the world they live in.
Since that is kind of essential to the core motivation of a wizard's intellect, insisting that a wizard cannot question, experiment, or attempt rational deduction to create new effects/spells, modify existing spells to have different effects by using slightly different conditionals, etc-- is straight up denying the ability to assume the role of a wizard.
Employing rule-zero because you dont want to try and work out how these experiments end in a more or less coherent way, and breaking that roleplay, is fucking lame.
LAME.
If you ask me, the better way to limit that kind of thing is to impose a cooldown, or material investment requirement to performing an experiment, and requiring controlled conditions-- then accepting whatever outcome the experiment yields as if it were a discovered rule of nature, with the same force as expecting a rock dropped from high up, to fall down and crash into the ground below.
Bullshit handwavy shit about "magic is SPEEECCIAL!" could totally apply to wishy-washy shit like sorcerers, but wizards approach magic like science. Such bullshit simply reeks when applied there.
Again, apologies, because this really is rather douchey, but is my honest opinion on the matter.