Neil,
ColorEyes Display is generally an excellent piece of software, and I have been using it for years with a Monaco Optix (X-Rite DPT94) colorimeter on Windows XP Professional. However, the most recent version of CED also has problems with DDC compliance. Integrated-Color is well aware of these issues and says as much on the front page of their website, much to their credit. Nonetheless, if DDC fails to work on one's display, it's a bit of a problem. The fall-back is using the on-display controls to set the necessary display input conditions for the ensuing profiling operation. This can be done, but depending on the display - it is awkward and not fully controllable. I've had this experience with my LaCie 321. Integrated-Color provides excellent support and they told me I've come close enough that my profile is good. Save for one minor issue I still have, they seem to be correct about this. So yes, my experience supports your recommendation, but there are issues, I-C is working on them and users can only hope they succeed; but it only takes one non-compliant link beyond their control in a rather complex chain for the whole DDC business to fail. The general problem afflicting DDC compliance has also been well-documented on the Mac Colorsync List-serve where Tom Lianza (if I remember correctly) pointed to what can only be characterized as a chaotic situation, given the huge number of video cards and displays out there and the absence of anything like an ISO document to help promote the kind of compatbility that would allow DDC to work regardless of the equipment and the O/S. This is a very weak link in the whole colour management chain, and one can only hope that the industry could fix it. But one gets the impression after reading enough about this stuff that there are hardware vendors in this industry who would prefer to frustrate transparency, compatibility and competition.
Turning to your calibration recommendations, I would suggest going easy on the cd/mm2. Working in a dimly illuminated environment as one should, 140 is likely to be too bright, causing the user to adjust the image too dark relative to the desired luminosity reflected by the print. I have found over the years that I've been using an LCD (with about 2 * 60 Watts of shaded room illumination roughly 10 feet away from me) that about 110 cd/mm2 provides a reliable display brightness relative to the prints. There's no hard and fast rule about this. It's trial and error.