I'm guessing this is in response to the necro buff?
General
- Non-self % based haste spells will no longer stack with each other.

It'll make stacking that with heroism worthless, but it's primarily being done to keep the new Power Infusion from being OP in conjunction with Heroism and talents like Icy Veins etc. as far as I can tell.

Here's what I've got so far...
Looking only at the number of actual attacks (excluding those within 45 seconds of procs)...
2.1 dual wield = 70 procs in 536 eligible attacks or 13.0597%
2.6 dual wield = 68 procs in 456 eligible attacks or 14.9123%
Warglaives++ = 74 procs in 473 eligible attacks or 15.6448%

Hmm, looks like we could use some testing with a faster weapon. Based on that information, it could either by flat ~15% proc rate, or something like 4 PPM.

Just grabbed a pair of 1.6 daggers, expect some data in an hour

edit:
Free file hosting by Savefile.com

Sorry if this has already been noted, but the Trouser's of the Scryer's Retainer is marked as using the Slayer's set in the Armor sheet.

Well, if 1.6 speed is 13.662%, and 2.6 speed is 14.9123%, I think we can rule out PPM. In terms of an initial estimate, I'd probably go with 15% proc rate and a 45 sec cooldown - it might not be exactly right, but it'll certainly be close enough to get the ranking right.

Just to correct something, the Scryer proc on the necklace isn't 100 crit rating, it's this. Seems to be considered a melee attack in regards to crit too.

I just noticed Rigid Bladestone and Delicate Fire Ruby have never been implemented in the game but are still present in the spreadsheet

Hmm... surprising. I can see making Aldor and Scryer do somewhat different effects, but I'm surprised that they're so totally different - seems like it'd be hard for them to balance. I almost wonder if it's a bug.

Anyway, if the proc effect is totally different, one can only assume that the proc rate might also be different; so until we can get some testing on the precise proc rate (once with a fast weapon and once with a slow weapon, anyway) we can't accurately model it in the sheet.

Assuming, however, that the proc rate is the same: the Aldor proc seems to be somewhat better than the Scryer proc - at least, given my gear, that's the case.

True. But as they're not getting recommended anyway, I'm not sure how much I care. There's still some chance that they'll be added, and I'd hate to take them out just to have to put them back in in a month or two. If they were being recommended as optimal, it'd be another story; but since the only harm they're doing right now is that you can pretend your gear is better than it actually is... I'm inclined to leave them (unless there's an overwhelming opinion I should do otherwise).

I can also confirm the spreadsheet blows up bigtime in Numbers (iWork)

Would be news to me that >[Dragonspine Trophy]< and >[Recipe: Haste Potion]< are:
a) percent based
b) non-self buffed

Don't see a problem there...

The change could have been in response to the necromancer buff, unholy frenzy...Thottbot World of Warcraft: Unholy Frenzy

Increases an ally's attack speed by 100% for 20 sec., but also inflicts 500 Nature damage to that ally every 2 sec.
(Stackable)

Though its clear it is directed toward Power Infusion+Heroism.

But blade flurry and slice and dice are both percentage based, so do they longer stack with heroism? Or am I misunderstanding what the patch note means?

Oscarvil, those are self-buffs. The patchnotes explicitly say
- non-selfbuffed
- percentage based
will not stack anymore.

Non-self % based haste spells will no longer stack with each other.

So unless you cant use blade flurry and slice and dice when silenced i would say it wont affect ya. It is for spells only by that wording. As stated above it was meant to nerf Bloodlust + Power Infusion. They choosed that opposed to choosing one of them (for example IV + Bloodlust were not stacking change which was replaced later with this rule).

I was actually pondering how hard it would be to write an AJAX version that can pull from Armory. And yes, I wasn't being serious about Numbers...

Armory pull is definitely a nice feature, but I've sort of decided that if I'm going to do that at all it's going to be in the aforementioned real-programming-language implementation of this rather than trying to do it from a spreadsheet. Basically, there's enough features (such as Armory pull, among many many others) that are enough easier to do from, for instance, Python that I really do feel that that's the correct long-term approach to making a fully-featured, easily-maintainable rogue calculation tool; hence, I've been working (on and off) on that rather than worrying about trying to hack Excel into doing what I want. But if you'd like to try to implement such a thing: by all means, feel free.

That sounds like a really good idea, especially since you can split it up into small maintainable classes. But unless you are going to fully utilize Python you might as well do it in HTML/ECMAScript and have it online IMHO.

Generally speaking, Blade Flurry and Slice and Dice are considered "spells" as well. They just aren't magic. Based on the patch notes I'd definitely say physical spells would in theory be affected by this change, although I am not aware of any non-magic haste buffs that are castable on other players. Since Blade Flurry and Slice and Dice are not castable on other players, they obviously are unaffected.

I don't really see any situation where rogues will be negatively affected by this change.

As has been mentioned several times now, the negative for rogues is heroism + necromancers in hyjal, and drums.

Drums are haste rating, not %-based; they are unaffected by this change.