It is just as I figured....

TP is always 100tp per second (TP/s) for every weapon of every type

So a delay of 4.2 would get you 420 TP per hit...
and a delay of 2.5 will get you 250 TP per hit....
So Each weapon will get you 1000TP in 10 seconds

So TP generation should NEVER be considered when talking about if one weapon is better than another, UNLESS it has "storeTP" or "Enhances Invigorate" on it...

There is what I like to call the "Divisibility of 10 phenomenon" though...
If 10/(Weapon delay) = (a whole number) then that weapon has a "Slight" advantage that doesn't matter if the fight lasts longer than 30 seconds, but only for a job that can't use invigorate

Example:

Cobalt Winglet has a delay of 2.6... it generates 100 TP/s. In theory, in 10 seconds it generates enough TP to start a combo. But since it's delay is 2.6, it will actually be 10.4 seconds before it has enough TP (Actually 1040tp) to do it's first combo.

Thormoen's Pride has a delay of 2.5 seconds. It also generates 100tp/s. Because it multiplies into 10 perfectly, you will be ready to unleash a combo at exactly 10 seconds.

Ifrit's battleaxe has a Delay of 4.3. It has the same 100tp/s, but after 2 hits its only at 860tp (Actually 862). It needs the third hit to get more than 1000 TP, which is actually 12.9 seconds.
So a weapon with a delay of 5, or even 10 would actually get a first combo off faster than the weapon with a delay of 4.3
A weapon with a 5 second delay would be ready for first weaponskill in exactly 10 seconds

Obviously this doesn't take into account how much TP you can generate from being hit... however, that is impossible to factor in... With block/Parry/evade and Stoneskin, it is possible to take no damage for the first ten seconds on many encounters

TP generated past 1000 is still available, so its not a big deal... just thought I'd mention it...