A debate.

Here's the RAND study they keep referring to.

"Legalizing the production and distribution of marijuana in California could cut the price of the drug by as much as 80 percent and increase consumption, according to a new study by the nonprofit RAND Corporation that examines many issues raised by proposals to legalize marijuana in the state...Based on an analysis of known production costs and surveys of the current price of marijuana, researchers suggest the untaxed retail price of high-quality marijuana could drop to as low as $38 per ounce compared to about $375 per ounce today." - I assume that's the good part for those interested.

This was the bad.
"While the state Board of Equalization has estimated taxing legal marijuana could raise more than $1 billion in revenue, the RAND study cautions that any potential revenue could be dramatically higher or lower based on a number of factors, including the level of taxation, the amount of tax evasion and the response by the federal government...RAND researchers caution there are many factors that make it difficult to accurately estimate revenue that might be generated by any tax on legal marijuana. The higher the tax, the greater the incentives would be for a gray market in marijuana to develop, researchers say." - More or less the same problem with tobacco or alcohol. Not an strong argument against legalisation that "it wouldn't make enough money". It's costing far more right now to enforce laws that penalize voluntary and hence victimless behavior. Not spending that money to investigate and enforce such laws and ultimately arrest or detain or convict people of criminal acts who have committed no violent or harmful actions seems like a fiscal gain even without the additional revenue that could be had.

It is possible that there would be negative price externalities (that is undesired social costs, such as employment and possibly driving) for marijuana use which are significant enough to desire some foolishness with the price mechanism in order to account for these. That is, this is an argument for taxing the stuff. What is not clear is whether a sufficient tax could be levied to account for it. It is very clear that a raised alcohol excise tax might capture many undesired social costs, in the same way that raised tobacco taxes will reduce casual use in particular and induce some heavier users (ie, addicts) to amend their social behavior as well. So an argument for a legal and accessible but regulated (such as you can't sell to teenagers) and taxed market should suffice.

In a similar vein, it would be desirable, regardless of the legality of the product, to design tests that could detect active impairment of the drug (ie, someone actually stoned) versus drugs that were still in the system from a weekend or even the night before. While at present, it suffices for police (and employer) purposes to detect any use whatsoever, this will undoubtedly not be true into the future and in many other countries, is already becoming less of an issue. An appropriate test for active intoxication and any impairment costs that provides may be needed at that time to appropriately penalize less responsible use of a narcotic or mind-altering substance. We won't get such a test without the cooperation of enforcement resources who would require them and we already cannot properly enforce such laws as driving under the influence without them. I fail to grasp how this as well would be a strong case against legalisation then.