I'm confident that if Theymos really wanted to upgrade bitcointalk he could easily do so in a much shorter time period.
That's correct but it doesn't make sense to introduce it before it has passed all of its tests
[it'll be a workaround that might introduce other problems as opposed to a solution].
As far as I can tell it's not being tested at all, not in a structured QA manner anyway, at least I'm not aware of any test plans or even requirements to be tested against. There are some volunteers putting their own time into this but that's not going to result in high (or even acceptable) quality software. I've already expressed my dismay numerous times at spending hundreds of bitcoins to create a piece of software and stopping 10-20 BTC short of pushing it over the finish line but what do I know.