The initial focus was on emails with juicy bits alluding to monkeying with data sets, avoiding FOIA requests, and blatant efforts to game the peer-review process to keep contrary views from being published. Shocking to some, but not me as I've been cynical about peer-review for a while now.
The deadliest finds came from programmers who started digging into all that code. They had a good road map to the worst parts in a log file from a programmer (Harry) who'd taken over the simulation from the original one. Unfortunately for the quality of the code these guys were climatologists who'd picked up some programming knowledge along the way, rather than being software engineers trained in how to handle this kind of very complex project. So they let a variable grow past the maximum value for its data type and then didn't understand why it became negative. Worse was their lack of any type of versioning or source control. Poor Harry wound up digging through files trying to find out what happened to the original data. At least some of it had been overwritten by "corrected" data . . . and the nature of the corrections wasn't always recorded. Some of the recorded ones are blatant manipulation. I'm on the record as being distrustful of simulation-based science and this is exactly why.
(Having spent time digging through old code myself I've got a lot of sympathy for Harry. I don't have any for his organization, which ran wild with the results of something they'd under-resourced and failed to check. My code wouldn't stand up to a mass review by Slashdot . . . but I'm not claiming moral superiority and a trillion-dollar tax code rewrite is justified based on my results.)
So there's no way to check the accuracy of those simulations without going back to the original weather station records (if they still exist), tree rings, and other raw data and starting over from scratch. Eric Raymond advocates doing just that as an Open Source project, making data and algorithms publicly available. That might be the only way to get something that we can trust. At this point any fresh analysis can't be done in an objective manner. There's too much money and too many careers resting on who's right and that's going to be an influence on every researcher who looks at the problem. Anybody not wanting to be forced into choosing a side is just going to avoid the whole field. Sure, we might get a martyr-type willing to be hated by both sides . . . but those are rarely the best analysts. An open-source project with partisans from both sides contributing would have a chance to develop a trustworthy model of our past climate. And that would make explicit the assumptions of the competing factions as they create models to simulate the future.
Once we settle the question of how much the planet is warming we'll have a chance to tackle other important questions:
1. Is the warming a product of human activity or just the normal fluctuation of an interglacial era?
2. What is the optimum average temperature for the world (Tongans and Siberians will disagree)?
3. How much damage will there be from being off that optimum, and who will suffer it?
4. How much will it cost to get us to the optimum (by restricting emissions or geoengineering)?
5. Is that expense worth it compared to other things we could do with the money?
6. Who's going to collect that money and oversee its spending?
The last is the most important. There's a lot of people fervent over global warming because they want "a global authority not answerable to voters" to be the answer to it.