To recap where we were: (i) using CEDP with this display is most likely not a great idea - high probability it won't allow DDC communication with the display.
I wouldn't say it's not a great idea. A more accurate statement would be that for an NEC that supports Spectraview II, it's not the optimal solution but CEDP combined with using the OSD to adjust the RGB channels would produce a decent result (better than Eye One Match or Datacolor's software and on par with BasICColor). My gripe is that if you don't already have CEDP, Spectraview is cheaper and will do a slightly better job since in addition to being able to hit the white point in the monitor LUT (which is essentially what you do when you adjust the RGB settings on the OSD), it can also perform other corrections and put them in the monitor LUT as well.
(2) If it cannot perform hardware calibration via DDC, he should avoid using OSD for this purpose (second best, fiddly solution) and use the $89 download of Spectraview II with his Spyder 2.
I would not advise using the Spyder 2. As for DDC none of the US NEC screens will work with any DDC solution other than SpectraView II. I don't know about other countries.
(3) If the profile verification results average at 1.0 or worse, he could do better by buying BasicColor Display and a more up-to-date colorimeter.
You could get a profile verification of 0.3 dE. Profile verification is not an extremely useful tool to judge the accuracy of a colorimeter because the results are relative to what the colorimeter reports and it's safe to assume that the colorimeter is going to agree with itself (unless it's say, a Huey which is extremely inconsistent... then it may not even agree with itself from readings taken a few minutes apart). A colorimeter merely measures input and spits out a number.
Let's say you have a Spyder 2 and your screen is outputting a particular color (and let's also say that this color is correct). The Spyder 2 measures that color and reports it as being incorrect so a correction is made in the profile. Upon verification it measures that color and finds it to be correct. The validation will pass even though it shouldn't have. Even with good colorimeters like the DTP-94, Eye One Display 2, and Spyder 3, this happens ALL the time. These devices are inaccurate when it comes to hitting the proper white point. A difference of 1800ºK between two pucks is not uncommon. Take a Spyder 3, an Eye One Display 2 and a DTP-94 and profile to 6500ºK and you'll get three different versions of 6500ºK. What's more important is that the colorimeter is reporting accurate color relative to the white point
and that is something that the Spyder 2 is not good at. It's an inconsistent device.
The original Eye One Display, Sypder 2, and Huey should all be avoided. The DTP-94 is an excellent colorimeter but doesn't do well with wide gamut displays (it would however be my first choice for a display with an sRGB gamut). The Eye One Display 2 and Spyder 3 are much better suited to wider gamut displays.