Do the clients in ad-hoc networks ever broadcast the SSID of the network like the access point does in an infrastructure network? I ask because I am trying to configure an ad-hoc network to be reasonably secure against casual hackers and I figure that the best way to keep it secure is to have noone know about it in the first place.
Ad hoc clients do not broadcast their SSID. However, they do respond to network probes (id requests) from other clients and applications. For example, Netstumbler will easily detect an ad hoc network using probe requests.
This is commonly called "security by obscurity" which is generally considered to be a marginal solution. For example, I often sniff with a spectrum analyzer, which will detect non-802.11 sources of 2.4Ghz RF.
If you need entertainment value, enter non-ASCII characters for the SSID. For example, Cisco uses a "x" in front of the two digit hex value of the character. I was using an SSID of x08x08x08x08x08x08 which is an SSID of 6ea backspace characters. Works fine with some clients and AP's that support Unicode. Blows up with XP SP2, most access points, Netstumbler, and Kismet. I've also used tabs, spaces, weird symbols, NUL's, and break chars, which cause various monitoring and management programs to hiccup. One would think that a program that can handles non-ASCII WEP and WPA keys, could easily handle a non-ASCII SSID, but apparently not. Such drivel will not stop a determined hacker (like me), but it will give the "casual" war a challenge.