I remember reading an article by Roberta Bragg some time ago that indicated she felt it was poor practice to name servers (particularly domain controllers) with easily identifiable names, such as Server1, DC2, etc.
In a recent Microsoft document (Medium_Business_Solution_ for_Core_Infrastructure), the advice is to do exactly that.
What do you think? Should a server's name reveal its function?
Not having obvious names for your servers is part of the idea behind "security through obscurity." I question the long-term benefit of giving your servers obfuscated names, however. For one, what is the productivity cost? Your administrators will spend time trying to type in server names correctly, and your users will probably never find the right machine they're looking for. (Of course
can help with this, but it's still a point.) Secondly, you still need to come up with a system for naming your machines, and in the event of a disaster or some other abnormal interruption in service, having clear machine names cuts down on a lot of head scratching.
If what your organization does is super-sensitive, I might buy into this theory. But for other companies, put the manpower into other lines of defense.