On Kelley:
I think Kelley "might" become a great pitcher (SP or RP). 
I don't bet my house on ANY pitcher with only 46 IP in the bigs.
Yes, I'm concerned about a major innings jump BECAUSE he was on the DL in '09.
I'm also concerned about a spur-of-the-Spring conversion, because if he stumbles as a starter and the club is FORCED to return him to the pen - you could easily end up with the Morrow-Morrow will his status change again tomorrow playbook being run again -- and it generally isn't a good idea to repeat the exact same mistake you just finished making.
I'm very, very, very worried about the bullpen - to the extent that I noticed the ONLY pitcher Aardsma has been better during the Spring is Olson.  (Okay - all hands for people wanting Olson eventually being called on to set-up and/or close.)
I don't get why the entirety of the Seattle web prescence went apoplectic about the Fig/Lop Flip-Flop because you either cannot possibly adjust from 2B to 3B in a month *OR* if you don't look like Brooks Robinson at 3B in the first week, you'll never improve.  But, starting and relieving don't require any kind of transition whatsoever.  (If SP and RP are the same - then 2B and 3B are identical, too -- it's just catching grounders and throwing to 1B, after all.)
The Ms currently have more pitchers than The Pottery Barn - (and about a gross of them already have MLB starting experience).  Opting to beam Kelley to the Bridge after a dozen innings in ST is the kind of knee-jerk spur-of-the-minute decision that likely does irreperable harm to the entire concept of players EARNING a spot in the lineup -- (because the next pitcher to have a good week with a 10/1 K/BB ratio can JUSTLY ask -- "So, why don't I get to start next week, too?")
If we're going to begin making rotation choices off of 12-inning samples, then shouldn't we justify making Tui the DH and cutting Junior based on their ST samples?  (Not like Griffey's contract is too large to eat, is it?)
Basically -- *IF* Kelley is as great as advertised -- let him dominate from the bullpen for the month of April to confirm the suspicion under GENUINE competitive conditions.  If, after that period, you are satisfied with the strength of the bullpen (in theory without Kelley), then you move him into a starting role in RESPONSE to other events - (SP getting hurt, suspended, rocked over multiple starts, etc.)
 

Busted a gut when I saw it - expecting: "Over-ruled" as the body.  :-)
Personally, I like the Santana approach for a number of reasons ...
Allows the kid to work on the new stuff "regularly" - w/o allowing multiple looks to the same hitters during the refinement (i.e. - oh CRAP) phase of adding a new pitch.
Naturally suppresses his year-long inning total -- 35 IP by July, followed by 80 in the second half is (IMO), a TON better than 80 IP in the first half - and then limited to 35-40 during your theorhetical playoff push.
Facing hitters once (9-hitter limit) is one of the reasons relievers have a better aggregate ERA than starters -- so you build his confidence with a better ERA -- but you give him 2-inning experience regularly, so he can adapt to the mid-inning warm-up cycles without expecting 27 hitters.
As a reliever, there will typically be someone else ready to get warm as soon as he enters a game -- so damage control on a bad day is quicker and easier.  The 8-run first disasters are a good way to destroy a kid's psyche before he's even started - (this is magnified by the importance of the game - which is why the same thing happening in AAA isn't as bad).