It's a fairly good list, though my qualm lies with this last part...mainly since an 'informed' decision today isn't the same as an 'informed' decision next week/month/quarter/etc. That's the hardest part about purchasing mining equipment. You may be making the best purchase (in terms of $/GH, or kWh/GH) in this very moment, but what's to stop Company X from developing 22nm (or 14nm, 10nm, etc) ASIC technology that makes your purchase obsolete?
You can make some fairly education assumptions. Miners generally are going to earn the bulk of their lifetime revenue front loaded. So the first 6 months is more important then the next 18 or even 36. You don't even need to worry about 22/20nm tech.
Here is why. 22/20nm is insanely expensive. Other than Intel (who doesn't use foundries they build their own fabs to their own specs),
nobody is shipping any product in 22nm (or 20nm). Not Apple, Samsung, AMD, Nvidia, nobody. Lets think about this for a second. These are companies with hundreds of ASIC engineers on the payroll, R&D budgets in the tens of millions of dollars who combined a couple billion cutting edge chips a year. 22/20nm is too expensive for them. Long before any Bitcoin ASIC company can look at 22/20nm any miner you buy today is going to be end of life anyways.
22/20nm will come but it is further away. The first (non-Intel) products moving into 22nm in late 2014 are .... smartphone chips. Why? Because the cost is sooo high right now it doesn't make sense. Would you buy an AMD graphics card with twice the computing power for five times the cost? I doubt it, just like you wouldn't buy a hyper efficient miner which costs 200% more than anyone else on the market. Cellphones are a unique industry where the product price is high, the processor is cheap and battery life (which 22/20nm helps) is everything. Even 2015 is dubious for 22/20nm. Just because a process node is available doesn't mean it is economical. It generally take 2-3 YEARS before the cost per transistor to fall below the prior node. This is just the cost per transistor, remember new chips means new multi-million dollar NRE. Unless a company can seriously undercut the competition that is huge risk to take. As for 14nm and below start worrying in 2020 or later.
There are a lot of risks to mining but if it helps you can ignore cutting edge tech.