Thank goodness for the bolded portion, I was starting to feel the pressure of some abject scrutiny... lol
There seems to be a little confusion (not unanimous, but general) regarding the recent advances made regarding latency and spellcasting in HB and the effects it will have. The HB Core recently had a property added to get the Spell Cast Time Remaining. Honestly it could have been there for awhile, I only found it because a user on a system with high latency asked if I could look into any form of pre-casting. I had a conversation with Nesox who told me about CanCast() and later while looking at HB changes and enjoying a fine Kentucky Bourbon the next thing I notice is the time properties... voila! The system latency time has been available for quite awhile, so with those two values you can now easily pre-cast and negate the latency impacting your DPS, HPS, TPS, etc. for non-instants. However, while a property for GCD Time Remaining was also added, it was not providing any additional value beyond the existing information. CanCast() uses these properties to allow the next cast to being while the current one is still finishing up (due to latency.) However, since the GCD time remaining isn't accurate, this means that melee (which cast primarily instant attacks) won't enjoy the same performance increase that caster specs do until the release with the GCD fix Apoc referred to is available. The result is that CC's (such as ShamWOW) that don't utilize CanCast() as the primary gate in determining whether a spell can be cast, are capable of realizing the performance benefits of eliminating latency by factoring in that time (via the newly added properties) in their individual GCD and Spell Cast wait logic. CC's that don't use CanCast() and don't use these time values, won't perform any better (or worse.)
There are some downsides to this improvement. Some CC's that cast instants which result in a buff or debuff without a cooldown (think instant HoT/DoT like Shaman's Earth Shield or a Warlock's Corruption) will result in duplicate casts if they depend upon the latency to cause a delay sufficient for the target to realize the buff/debuff. I encountered this in testing with ShamWOW in the latency improvement release and addressed ensuring Earth Shield isn't cast more than once per second. Apoc confirmed it will potentially hit other CC's implemented for classes having spells meeting this criteria dependent upon their implementation. So while you may realize a spell casting throughput increase, the CC may require an update to avoid a DPS/HPS loss.
Now speaking specifically about ShamWOW, the design objective is to implement ability priorities suitable for all facets of game play giving equal importance to leveling, pvp, and raid environments for ranged dps, melee dps, and healing. So while it is intended to be extremely good at leveling or PVP, it's also designed to be good at raiding (one reason I wrote LazyRaider.) Ultimately a human should be able to beat HB in combat scenarios involving movement or that benefit from preparation (for ex: raid strategy to move to certain spot, casting a Shock so Focused Insight buff is available prior to needing it so GCD isn't wasted, etc.) I GREATLY appreciate those that have enjoyed using ShamWOW and recognize the time and effort that went into it, but in the end a L337 player should outperform it in anything other than a stationary fight against a Test Dummy, and should have beaten it handily at that before the latency api's impacting your spec are fully added to HB.
Hope that clears things up, or at a minimum doesn't muddy them further,
Bobby53
)