Thomas,

Your suggestion put me on the right track! There is no diskette drive, but there is a CD drive. The problem is not the CD drive per se, but it is an issue with the long abandoned framework I chose to use back in the "dim time". The framework sets the path to include the source directories for convenience while debugging on the development machine, On my dev machine that is the E: drive. After rebuilding the troublesome workstation with a secondary data partition on the new and larger hard drive, the CD drive became E:. Changing the CD to F: so E: was undefined solved the problem for now.

Your mention of the "optical IO" caused me to examine the failing "CreateFile" operations more closely and I noticed the ones on the C: drive completed very quickly while the ones on the E: drive took incredibly long. The CD light never lights, but XP must be polling the drive anyway to see if a disk slipped into it unnoticed.

Now the only question would be "Why the difference in where in the logon process I start the program?" I don't plan to pursue that question, but I am curious. I suppose the good thing is I learned how to use ProcMon to help troubleshoot the issue.

Now I"ll have to waste a day figuring out how to fix the near-dead framework. ("Was you ever bit by a dead bee?" ... Walter Brennan)

While you didn't exactly solve the issue, I'll give you credit for more than a simple "help". It might have taken me another couple days to notice the drive reference as I had already overlooked it for a couple days.

Thanks for the assist
...Jim

>Perhaps try to disconnect all slow disc-similar devices like optical IO or diskette drives
>by removing the power chord. very low probability, but as neither network nor AV is in the game
>the typical suspects are gone.
>
>...
>
>regards
>
>thomas