Teergrubbing (wikipedia) is the act of stealing resources from spammers so their spambots become less effective.
When spambots update a page on a wiki they probably don't wait for the response, so we cannot engage them immediately - we will need to remember their ip-number in a blacklist.
When they revisit at a later time we can start to pester them in two ways:
- keep the connection open as long as possible (days) since connections are a limited (but not small) resource on the spamming machine. If a spambot has a limited set of threads we put pressure on that resource as well.
- very slowly feed an endless stream of data into the connection, this will require the bot to allocate more and more resources for page buffer. If a spambot heuristically scans for text entry fields in a html page we might want to hand it an endless set of form fields (it has to remember them all for in a POST command, one of them might contain a session number) but no text entry fields. We have to do this slowly so we don't saturate our own internet connection (which could be quite costly too).