Both have their merits..
The admin script is good when you need to get it done on as many systems as possible RIGHT NOW!!! My models of this process remove the successful system names from the source, allowing me to repeat the task until all (mostly) are complete.
The login-script based model is good when you want the most thorough coverage. It can be problematic for remote users, particularly those that work at home, sleep/hibernate, then "wake up" at work - never running the login script. I prever to use a cattle prod / bat on those users, but it's usally the CTO/CIO (who should freakin' know better!) that's guilty of this.
My login script based tool provides an API for the login process - you write the detection code and use the APIs to send the result to the central server. The code on the back end is small - a single Kix-based Windows service at 20K, a config file, and whatever backend scripts you write to respond to the requests. The API library is only 5K.
Glenn
_________________________
Actually I
am a Rocket Scientist!
