A Bot is a common name for a software agent. A Bot interacts with other network services intended for people as if it were a real person. When communication between two people is mediated by a computer, a Bot can replace a human or either end. A Bot can replace the human to various extents as discussed under the Turing test.

It is quite common to believe that Bots are created to harm humans. This is not necessarily so. As anything out there, bots can be used by humans to do good or evil, because the Bots do not have mind of their own.

At the same time Bots represent a very interesting area of software engineering called artificial intelligence (AI). Developing a good Bot is quite challenging and it is fun. There are also specific algorithm and programming models used to develop Bots. It seems that Bots have many of the same "parts", namely:

• a host - a runtime environment where Bot lives; can be Java, Python, a Web browser, etc.
• a stimulus and response protocol - the search Bot has the search string as stimulus and the results table as a response; the chat Bot has a text string as the stimulus, and the text string as a response
• a knowledgebase - a database that holds the responses to the standard situations Bot can encounter
• a history - the last N of its own actions for evaluating next actions; for the web site crawling Bot the history is a lists of already visited links
• a decision-making heuristic – an algorithm that decides what to do next upon stimulus

If interested, check out my lst of distinguished bots.

The known spells against evil Bots include some sort of challenge-response system. If you look at the list of common Bot body-parts above, it's the specific stimulus and response protocol that is being used to defeat against a Bot. Other parts of a Bot seem to be useless and inconclusive in telling a Bot from a human (at least at the moment). If interested, check out my list of spells against evil bots.