You and I must play Guardian tank a bit differently. I've never felt like my CD's "gnashed", & I use cyclone slash for generating AoE threat on multiple targets as I make the pull. Not saying I'm playing it well, but I don't feel like I suck at it.
What I was talking about with going to a large pool, is that the focus generating part of the string would vary moar better over the long term, and by tweaking the ability cost the string goes more solidly from
A>B>C>A>B>C>A>B>C>A>B>C>A>B>C>A ~ to ~ E>A>B>C>A>B>C>D>E>A>B>C>A>B>C>D>
A & E being focus generating abilities B,C,D being GCDs (1,2,1,2..... Vs. 2,2,1,3,2,2,1,3)
As it sits right now, if you stick with a stable rotation you either starve or "max out". So a resource nuetral string for Guardian involves what I'd call mini burn phases. It's got a weird "incomplete" feeling.
This brings up some important points -
1) if you tweak the CDs for the defense spec - you wreck the vigilance spec. - but you have identified what makes the Vigilance spec fun to play. And it's that it gives the player options without messing up their play waiting on CDs. They can keep a string of casts tight, without actually planning ahead ~ it's a twitch gamer's ideal spec.
2) people can use these abilties differently, sometimes wildly so.
3) we all like our choices
4) an "intended attack string" - is what ended up making Tk Sage & Gunnery Commando boring. Because if an ideal string exists, much deviation sacrifices too much dps. Abilities on proc add some very interesting qualities to game play, but the proc still needs to let players "roam". I kind of feel like the 10 second duration of some procs is too short, but then again a longer duration might be too easy. If a proc is activated & consumed once every 10 seconds, it eats at least 30% of those 10 seconds.
5) Multiples of the GCD are relevant when setting CDs, casting times, durations, channel lengths. I never really thought about why the 2 piece set bonus for a 1.5 second reduction in HT CD was chosen ~ it's because you wouldn't notice anything less, and 3 seconds would have been impossible to balance.
Now I understand why you are so interested in going at it from a "CD first" perspective.
But I still say CD & casting time are being used to accomplish the same goal - the difference being that casting time can be influenced by alacrity, interrupts, and pushback. A small amount of alacrity may not influence a orderly string of abilities at all, but enough alacrity can create space enough for an extra ability to be cast out of turn. But casting times break the rigid structure to a degree. Mind Crush for example, doesn't interchange well with any other ability in the Sage toolbox, except for maybe Salvation. But neither of those abilites fit particularly well anywhere, and Salvation on rotation is a very boring way to heal. While Weaken Mind interchanges quite well, and even has a 12 second duration ~ for optimal refresh timing.
It's pretty obvious that BW thought the way to go about it was to make sure that a class always has abilities that share the same CD or other abilites whose CDs total the longer CD, so if nothing else, they have forks in the road. Too many long CDs - creates a bursty class, that must spam a basic attack. These are also difficult to balance for PvP, because of their potent "opener". While too many different CDs creates a class with no real rythem to it. Which is why we see the multiple of 3 repeated so frequently with abilities. It's also why many debuffs & buffs on proc work in multiples that don't divide by 3 ~ essentially forcing players to refresh early, less than optimally, or let them expire.
Other multiples of 1.5 give you 4.5, 7.5, 9, 10.5, 13.5, 18, 22.5, 27, 31.5, 36 ~ seems like I've seen 4.5, 7.5, 9, & 18 in the game, as for the rest... I'm not sure. Any of those could be used to design abilities that played differently. But an ability that would balance with a weird "off GCD" cast time (such as 1.6, 3.1, 4.6, 6.1), could easily become imbalanced by alacrity.
Imagine a pair of 2,4,6 second dps casts as a class' abilities - add the talents to reduce the casting times to 1.6,3.3, 5 & then some alacrity gear to make those 1.5, 3, 4.5 second casts ~ the same abilities become real jaw breakers that are used in rotations with instant casts in a completely new way. A few talents could be used to make 3 specs in the same AC require completely different gear. It could be what makes the distinction between a heal spec and a dps spec for another example. Without the alacrity, the pair of abilites can be used efficiently and interchangeably in strings only at every fork- but with the right talents and alacrity... they could fit right anywhere. Power/Alacrity Vs. Crit/ Surge. I'd be surprised if this isn't already in the game somewhere, on one of the pure dps AC's I havn't played.
I Example of the base ability:
1) a 2 second casts fits tightly as part of a 10 second string, with any class design that has 10 second procs or debuffs.
2>1.5>1.5>2>1.5>1.5> ~ or ~ 2>1.5>2>1.5>1.5>1.5>
2) a 4 second casts also fit well 4>1.5>1.5>1.5>1.5> ~ or ~ 4>2>4> ~ or ~ 4>2>2>1.5> ~ or 4>1.5>4>
3) a 6 second cast ~ really only fits in once per 10 second string 6>2>2> ~or ~ 6>4 ~ or ~ 6>2>1.5>
*9 options that make sense
IIOn alacrity/power spec these same abilites become interchangeable with instants at the expense of creating the tightest possible strings that are just right to refresh 10 second procs. Tthe spec also gains an extra GCD worth of space in it's string roughly every 15 seconds, at the expense of passive resource regen during these abilities.
1) 1.5>1.5>1.5>1.5>1.5>1.5> = 9 seconds
2) 3>1.5>1.5>3> ~or~ 3>1.5>1.5>1.5>1.5> = 9 seconds
3) 4.5>4.5 ~ or ~ 4.5>3>1.5> ~or~ 4.5>1.5>1.5>1.5 = 9 seconds
*6 options that make sense
IIIif it's only the 4 second cast that benefits from the alacrity talents & stats:
1) 2>1.5>1.5>2>1.5>1.5> ~ or ~ 2>1.5>2>1.5>1.5>1.5>
2) 3>1.5>1.5>3> ~or~ 3>1.5>1.5>1.5>1.5> = 9 seconds / 3>1.5>3>2> ~ or ~ 3>3>1.5>2 ~ or ~ 3>3>2>1.5 = 9.5 seconds
3>2>3>2> ~ or ~ 3>3>2>2 = 10 seconds
3) 6>3> ~ or 6>1.5>1.5> = 9 seconds / 6>2>1.5> ~ or ~ 6>1.5>2> = 9.5 seconds / 6>2>2> = 10 seconds
*14 options that make sense
It's an overly simplified example, since it assumes the the ability that generates the proc is cast in the same "position" every 10 seconds, But ~ which of these looks like more fun to play?
0.5 seconds of time can dissapear over 10 seconds of casts, just due to the tiniest hesitation between abilities. Pushback can also eat up casting time. 12 second procs would work tighly for the same abilities in either spec. In play "perfect 10's or 12's" may actually be unplayable with less than 11% alacrity. This could actually be used as a mechanism to bar the alacrity/ power spec from using the procs ~ they could be reserved for the crit/surge spec. The alacrity stat required could be tuned down to a level like 5%, so that some of the effects became a reality while leveling, or so that a hybrid build/gearing became viable with endgame gear.