Some quick background: played a LOT of D2 (at least 8 level 75+ chars that I can remember offhand and 2 with best possible gear) and WoW, so I have a lot of experience as to how Blizzard balances their games and abilities.
First, there seems to be a lot of confusion over how the wizard plays solo. Everyone has a chance to play with someone else, as we all experienced with the templar. On a personal level, I feel the templar is the perfect merc for the wizard, as he's more of a tank than a damage dealer, which complements the wizard well. The templar will definitely die to packs of strong enemies (like champions) if not assisted quickly, but he should be able to solo single mobs that come after you or tank groups for a short period of time as long as you maintain his gear. With this in mind, high AoE dps could be considered a form of crowd control as it would allow your templar to take the agro while you burn mobs down before he dies. Playing a sorceress in D2, this held true even in hell, as I could allow the act 2 hero to run in and blizzard on top of him while using glacial spike/frozen orb. As long as my dps was high, he wouldn't die. I plan on using both blizzard and comet to keep up high AoE damage on him.
Second, there are a lot of assumptions being floated around based on the beta. To rate the effectiveness of a spell based on part of one act in normal is a dangerous mistake to make, as most people who played D2 through hell can attest. I don't know how many times I had to scrap a toon because a spell I thought was awesome in normal turned out to be too underpowered in hell. Until we actually know how hard inferno will be, all we have is speculation. Speculation is okay, as long as everyone takes it with a grain of salt.
Third, a fair number of people have been comparing straight spell dps and picking winners based on that. Ray of frost is a perfect example; in paper, ray of frost will outperform all other spells at sustained single-target dps. The problem with this (which fortunately other posters usually address) is that twofold. First the spell is being examined in a vacuum. Comparing two similar spells without accounting for other spells and effects is flawed logic. The burden of channeling RoF is that it either hinders your ability to rotate other spells in, or hinders the dps of the spell itself. Like everything, it's situational, but the more you rotate in other spells, the less overall damage you're getting from ray of frost. The second problem is that RoF's high dps assumes that you don't need to be mobile. When playing with other people (or using the templar merc), it's reasonable to assume that you will get a decent amount of time to channel the spell. However, I'm sure there will be encounters where the wizard must remain mobile for a large portion of the fight, during which RoF will become next to useless. This is not to say that ray of frost is a bad spell; I will be running with it most of the time and we don't really have other options for a single target AP dump. It's just dangerous to look at dps to assume spell superiority.
Lastly, cooldown reduction is far less straightforward than many people make it out to be. Frost nova is a great example: if you only need one frost nova per encounter, and don't have anymore than one encounter requiring frost nova every 12 seconds, a cooldown reduction is 100% useless. However, if either of those conditions is not true, a cooldown reduction can be far more effective than 30%, or whatever cooldown reduction is used. This will require testing: how often are you hitting a spell button shortly before the cooldown is finished? Does cooldown reduction significantly change this? Is there a way to modify your playstyle to not require a cooldown reduction? Cooldown reduction in D3 really seems to take a lot of potential power away from spells, and I personally believe should be avoided when possible.
In regards to evocation vs. critical mass, I expect (based on how Blizzard has balanced such things in the past) that critical mass will result in a larger reduction in cooldowns, at the cost of consistency. There will most likely be an ICD, as to not have one would make this ability scale very well with gear and force many wizards who wish to optimize their dps to stack crit. It IS possible that it does not have an ICD, as it is obtained at level 50, but I wouldn't hold out much hope.
In keeping with my own standard, this is mostly speculation, albeit informed speculation. I welcome comments and criticism.
Wizard should almost always have a tank during actual play.
Very little known about effectiveness of spells in inferno.
Can't compare spells with different mechanics in a vacuum.
Cooldown reduction: situational, not generalized.