That's a great observation. Still, I've heard that SSD caching is kind of a stopgap measure itself, until SSD becomes cheaper. If you already have a 256 GB SSD (like the ones most of these people are waiting for), what's the point of SSD caching? Meaning, isn't the point of SSD caching is that you can do it with smaller (40/60 gb) SSDs, not the "large" 256gb models? Inquiring minds would like to know.
Of course SSD caching is a stopgap measure. The only way to get a 2TB SSD otherwise is to RAID0 four 512GB drives together, which is about $5,000-$6,000 and shoots the MTBF clear in the foot (and having nothing but enterprise customers, SSD lifetimes are much shorter than we'd like to see, and that includes even FusionIO).
However, for most users and most use cases, it's "good enough". I have a Momentus XT in this Mini, and in my Thunderbolt MBP15 -- boots, app launches? As fast as an SSD equipped laptop. It won't shine when running database queries and doing DW work since the test databases aren't in the 4GB cache, but that can be resolved with a larger cache (256GB instead of 4GB). The difference in boot times between running off a Momentus XT and an actual SSD is < 1 second. It approaches zero. Yes, the SSD is still technically faster, but barely. Same with level loading times on the Mini. And this is on a tiny 4GB cache!
Again, it'd probably be an option. For most users, caching is "good enough" -- way faster for
everything they do. Only for the minority would users want to ensure that certain data is always on the SSD, and for them, caching can be turned off and people can manually partition their data (partitioning being something the average user doesn't want to bother with), creating symbolic links, etc.
Still, ship dates are basically post-WWDC (6th through 10th), which is interesting.