L1-A and security

[De Clarke Sys Mgr]

This may be just naivety on my part, but...

(1) I don't want my users using single-user mode to do silly or
    malicious things

(2) I don't want to be called in the middle of dinner to "please come
    and type in the root password because my workstation just crashed"

(3) I might settle for having sgl-user mode prompt for a username and
    password, checking it against its passwd file;  then the user could
    log in, but their name and the timestamp would be recorded (where,
    I wonder, that they couldn't gp and erase the record?).  Anyway, you
    get the idea.  The point is that strangers are more likely to be
    malicious than my own small, well-known user group;  and within
    my small group, people are much less likely to do anything stupid
    or "humorous" if it can be traced back to them.

This is obviously a blue-sky notion.  Can anyone point out either why it's
absolutely ridiculous or (if it isn't) how it could be implemented?

..............................
De Clarke, Systems Manager, UCO/Lick Observatory
Still trying to teach it epistemology...

root at helios.ucsc.edu, postmaster at portal.bitnet
voice: 408-429-2630   fax: 408-429-2730
The usual disclaimers apply;  your actual mileage may vary.