OP updated for both changes.
PoJ, though weakened, retains a slight edge for PvE. 15% always and 25%+ for most of a fight (since we should commonly be sitting on 2 HP). It should average better than Long Arm's 45% for 3 out of (less than) 6 seconds (average 22.5%), without Long Arm's additional issue of fast/slow/fast/slow jerkiness (or some crazy Silence effect). On the parity level, the Druid talent was likewise nerfed, so we're not being unfairly discriminated.
SoT vs SoR redux.
SoT is 14% Weapon + Censure.
SoR is now 7% Weapon.
Calcs:

←
Click Here
Assumed: Good old Normal Starshatter (3.6 speed, 12055-18084 damage), and 30k AP. Also assumed that SoT still is not normalized while SoR is still normalized. Why is one normalized and the other is not? Don't ask me, I just know what we've found based on empirical testing in Cata and before.
SoR
((((AP/14) * 3.3 ) + Weapon damage) * 7%)
((((30000/14) * 3.3 ) + 15070) * 7%)
((2142.8 * 3.3) + 15070) * 7%
(7071.4 + 15070) * 7%
22141.4 * 7%
1550 per target
SoT - Lots of data copied from Immediate Truth number crunching
(((AP/14) * Weaponspeed ) + Weapon damage) * 14%
(((30000/14) * 3.6) + 15070) * 14%
((2142.86 * 3.6) + 15070) * 14%
(7714.29 + 15070) * 14%
22784.29 * 14%
3189.8 per hit
15 seconds / 3.6 swingspeed = 4.166 autoattack in 15 seconds
Manage a CS every 4.5 seconds, so 3.33 CS in 15 seconds
And claim a TV every 4th GCD (implausible, but we'll say it's true), so 2.5 TV in 15 seconds
Combined 9.996 attacks, we'll round to 10
31898 in 15 seconds
Formula for single stack Censure
126 + 11% SPower per 15 seconds at 1 stack
where SP equals AP/2.
126 + 5.5% AP over 15 seconds at 1 stack
126 + (30000 * 0.055) over 15 seconds
126 + 1650 over 15 seconds
1776 over 15 seconds
355.2 per stack per 3-sec tick
We'll say first tick is a 2 stack, every other should be a 5 stack.
The number beside SoR is targets. Time is when (unhasted) melee attacks would hit.
| Time | SoR 3 | SoR 4 | SoR 5 | SoT + Censure | SoT Combined |
|---|
| 0 | 4650 | 6200 | 7750 | 3190 + 0 | 3190 | | 3.6 | 9300 | 12400 | 15500 | 6380 + 710 | 7090 | | 7.2 | 13950 | 18600 | 23250 | 9570 + 2486 | 12056 | | 10.8 | 18600 | 24800 | 31000 | 12760 + 4262 | 17022 | | 14.4 | 23250 | 31000 | 38750 | 15950 + 6038 | 21988 | | 18 | 27900 | 37200 | 46500 | 19140 + 7814 | 26954 |
|
Results:
SoR is winning at 3 targets, though the margin is getting rather narrow (less than 1000 damage) by 18 seconds.
GCD for seal swaps would likely be net loss for less than 4 or 5 targets. Transitions with dead time (boss unhittable and adds not yet spawned/in range) would be perfect times to burn GCD on seal swaps.
A much stronger weapon without barely any extra AP skews slightly towards SoR (since Censure doesn't benefit from weapon damage), but it's unlikely to every be significant enough to skew results to fewer targets for SoT, so bottom line should be correct for all levels of gearing.
Bottom Line: If you have no free GCD then stick with SoT unless you have 4+ targets that will live for 30ish seconds, chop off 5 seconds for each target above 5 (so 25 seconds for 5 targets, etc). If you have empty GCD while waiting for adds then swapping to SoR at 3 targets is a gain. Thus, in general, 3 targets is full AOE territory (SoR, HotR, DS).
Standard caveat - numbers crunched at present values of abilities. Things could always change (again) tomorrow.