Blocking works for chats originating from website chat box and texting. From the blocked guest's perspective, the operator will simply stop responding even if new chats are initiated.
A block persists for a period of 24 hours. 24 hours will be long enough to discourage most abusive guests, but it's not so long as to present a serious long-term problem for dynamically-allocated IP addresses. We usually don't want to block IPs for really long periods of time because they can be redistributed to other users. And of course, there are libraries with public computers available, where multiple guests will use the same IP address in a 24 hour period. If you work in a library with public computers, you'll probably want to work out internally how you handle that. For example, there might be an internal policy to check an IP against a list of public PCs before blocking it.
If this doesn't do the trick because the guest is super persistent, contact us and we can see about blocking that problem guest for a longer period of time. We can also remove IP blocks before the 24 hours if over if needed.
If you are getting spam from bots on an IM network (e.g. AIM), you can log in directly to your gateway'd account to block that username permanently.
Operators using an external client such as Pidgin should click on the "chat management link" that displays at the top of each chat and begins, "For transfers, etc...". That link takes you to a page that looks something like this: