I apologize for the length & math, for a summary just skip down to conclusions. Some (all?) of this may even be old information I was just unaware of. :)

I have never been quite happy with how the weapon DPS was calculated in the item formula. Neither was Hyzenthlei when he posted the original values in the ‘From Epics to Errors’ thread:
I only have an old spreadsheet, so my information is based on numbers and comments on that spreadsheet. Whether someone has further developed the DPS calculation I don’t know, but for example http://www.wowwiki.com/Formulas:Item_Values doesn’t mention DPS calculation at all.

The values for weapon DPS are indeed messy. You can fit 3rd order polynomial to pretty much any data. Now the fit is fine for the data it was made with, but it already breaks when trying to extrapolate say DPS of the weapons from Kel’Thuzad. There are other quirks too, like DPS of uncommon 1H weapons starting to go down after iLvl 85.

A new 3rd order polynomial fit fixes the problem with the new high iLvl weapons, but it can be expected to break down as soon as iLvls get significantly higher again. Unfortunately making a simpler fit seems very hard. An exponential growth fits the high level data quite well, but at 4 parameters it’s equally complex to the 3rd order polynomial and not necessarily any more robust for extrapolation.

To study the issue I looked up the minimum and maximum DPS of uncommon, rare and epic weapons for each iLvl (disregarding obvious anomalies like caster stat DPS reduction and elemental damage). Then I took a mean of the min. and max. DPS and further averaged over 3 levels to reduce any spikes. This data is plotted here for 1H and 2H weapons of different qualities:

Some of the graphs look quite similar and I found a couple of things that surprised me at least. Here’s what the same data looks like if 1H DPS is multiplied by 1.3, rare items are shifted 6 iLvls higher and epic items are shifted 12 iLvls higher:

I still haven’t had success with a better fit for the actual DPS calculation, but it seems that the DPS for all 1H and 2H weapons is derived from the same function with a few parameter changes. There is one obvious bump around level 40-45, and a smaller one at lower levels. Other than that the increase in DPS gets steadily faster.

Conclusions:
An epic weapon has DPS equal to a rare weapon six levels higher. A rare weapon has DPS equal to an uncommon weapon six levels higher.

A two-handed weapon has DPS equal to 1.3 times the DPS of a same level one-handed weapon.

It would be useful to find out the actual formula from which the DPS is calculated, or at least get a simpler fit so high iLvl weapons could be more accurately predicted. Good ideas are welcome.