I like your analysis. There are a few other aspects that determine my views of a coin. Mainly, mining.
I do not think that 100% POS coins are viable. Whoever owns them at the start, becomes "rich". The only way for the currency to spread is for the original owners to spend/trade it. These coins become sort of a cartel; closely held by a few key players, none of whom can really trust each other not to devalue their holdings by selling them all off. So they all have to slowly dispense their coins in order to maintain their "value" but not too slowly, because the market could crash at any time if one of the cartel members decides to liquidate. A classic Mexican Standoff. With no new coins being generated by outside parties, any "new blood" has to buy into it by trading something that has an extrinsic value, i.e. a value which is not determined by a cartel. NXT falls into this category.
I think that a POS/POW combination system such as Peercoin is really interesting, and could have a viable place in the future crypto marketplace; or perhaps POS would simply be adopted by Bitcoin and others as an extra way of generating coins once mining income falls away.
I think that coins with alternative uses such as DNS (Namecoin) also carve out a viable niche. Apparently this is the kind of thing that BitShares is going to get into, so I am excited to see where they go with their ideas. Proof of ownership in general becomes an interesting idea. I read an article suggesting that copyright and licensing could be integrated into a coin-like, peer-to-peer structure. It would certainly be interesting if, for example, a computer game's DRM could be based on peer-to-peer software, and licenses for games could simply be "coins" traded in a marketplace; whoever shows up in the ledger as the "owner" of that game gets to play it. But that ledger is not held by the developer, and is not subject to a single point of failure, e.g. Sony's servers go down and nobody gets to play their games for a few hours.
Of course this idea can be generalized to forms of tangible property as well; the idea of a public ledger denoting all transactions is found in every County Courthouse in the US, and the only way to determine 100% ownership of a piece of real property is to follow the ledger from the founding of the county, to the present day. This is how title companies make their money, by maintaining their own ledgers (and researching from the public ledger when necessary) in order to provide a guarantee to a buyer that the seller actually owns the property. It's kind of hilarious to me now that I've made this connection; Bitcoin is based on a centuries-old idea for proving ownership.
Other alternative uses include scientific computing enterprises such as Primecoin and Riecoin. Hopefully this will be expanded to things such as Folding@Home and other similar endeavors. People are already volunteering their computing resources for these purposes; it would only be fair to give them something in return, and it would be almost trivial to adapt an existing scientific computing platform into a coin "skin".
Now getting back to the mining bit... I think that the way that Bitcoin was/is mined/inflated is really a great model. Maybe the most perfect implementation that we have. Popular acceptance of a coin is to some extent based on the idea that it is "fair", that it is not extremely concentrated in the hands of a few individuals. And I think that we have to face the fact that the first "popularity" that needs to be gained is within the cryptocoin community itself, which has a large proportion of miners. And a "fair" coin has a lot greater chance of becoming popular among miners.
So, what determines "fairness"?
(1) Choosing an algorithm that ideally is not CPU-only (which just gives botnets and AWS farmers a leg-up) but is also not tied to existing ASICs (so, at this point, not SHA256). GPU-targeted mining is the way to win the hearts and minds of small miners, who IMHO are the "long tail" of the crypto community.... they may not have as much hashing power as the big guys; they aren't millionaires, and they know they're not going to become millionaires by mining (but they are non-trivially supplementing their normal incomes); but they are extremely numerous, fairly well-informed, and vocal. These are the folks who form the initial community around a coin, and they can help or hinder the marketing aspects of the coin to a large degree.
(2) No or extremely small pre-mine.
(3) Little chance of an "insta-mine", where difficulty is absurdly low for the beginning period and so much power comes to bear on the network that blocks are being generated every half-second, with tons of orphaned blocks. Setting an appropriate initial difficulty level, and an appropriate block reward, and an appropriate difficulty-scaling algorithm are all very important here.
(4) A fair and open launch, meaning: wallet, mining, and pooling software that functions properly at launch.
(5) A chance to mine over a fairly long period of time. If 90% of coins are mined within the first 3 months after a coin's launch, you are going to have a hard time persuading miners to support the coin over the years to come. A community can take some time to develop, and I think that a fairly high mining-inflation rate over the first several years is perfectly reasonable. Again, I think that the BTC model is pretty darn good.
Riecoin did something interesting in its recent launch; the block reward for the first 500 or so blocks was 0. This allowed all interested miners the chance to get up to speed, to make sure their wallet and mining software was working, etc. The difficulty scaled normally during those first 500 blocks, so by the time actual coins started being produced, the difficulty was at an appropriate level for the network hashrate.
Bitcoin went through all the stages; CPU mining, GPU mining, FPGA's, now ASICs. It had first-mover advantage, which is huge. At this point, with the fiat investments put into BTC, I don't see it ever losing its lead. We have "real-world" investment instruments based on BTC (private funds already exist, with the Winklevoss ETF supposedly coming soon), so I don't think there's any way to stop it now.
However, I do see there being a permanent place in the marketplace for niche coins, as explained above (such as Peercoin, Namecoin, etc.) and for different "denominations" of plain-jane coins (which is what I call coins with no notable added features). So if BTC is the hundred dollar bills, LTC is the singles, and DOGE is the pennies. Theoretically, this should not happen; after all, BTC is just some digits in a ledger, and it can always be sub-divided into smaller and smaller amounts.
But at the moment, the BTC transaction fee is a show-stopper for microtransactions, where DOGE seems to have found a good foothold. Nobody wants to pay a $0.06 transaction fee in order to send another person $0.10. But that is where BTC stands today. (And for good reason; preventing spam in the blockchain is, of course, a necessary step.) But as long as there is some demand for a level of payment which would be considered spam in the BTC blockchain, there will be a place for a plain-jane, much-lower-valued cryptocurrency such as DOGE.