Introduction

   Currently, many ANTISPAM rule implementations exist in several application types, which range from SMTP servers to clients for end-user e-mail readout.

   Such implementations, regardless of their good or bad quality (and they are usually excellent), suffer eventually from a serious drawback: they can only run in a specific application (MTA software). They are not easily portable to other applications and, in most of them, the local user can't modify the antispam behavior for his e-mail address without need administrator privileges and without affect others local users.

   For that reason, sharing ANTISPAM rules becomes difficult (quite often, it is not trivial). As a consequence, some applications present very good ANTISPAM features, while other applications do not have such good features but present other excellent features that are complimentary to those of the former applications.

   Although much effort is made to implement ANTISPAM rules, most of this effort does not substantially contribute to the welfare of the community. At the same time, the production of SPAM (unwanted e-mail) keeps increasing at an astonishing rate.

   General purpose libraries are available, which support functionalities that are used by a huge number of applications. An example is ZLib (general purpose data compression library). It is therefore natural to ask: can't we develop a general purpose library for the implementation of ANTISPAM rules, so that these rules can easily be shared by all applications that handle e-mail either in transmission or readout?

   This is the goal of LibAntispam: becoming a general purpose library for the unified implementation of ANTISPAM rules and procedures. These rules and procedures can be shared by all applications that are in charge (directly or not) of the transmission of electronic messages under SMTP protocol. With this library, network administrators can share ANTISPAM rules more easily and dynamically, not worrying about which SMTP server their colleagues have been working with (Sendmail, MeTA1, Postfix, Exim, etc.). The use of LibAntispam unifies the efforts of different people, focusing these efforts on the reduction of the amount of SPAM.

   The main goals of LibAntispam features are as follows:

Their use in any SMTP server design should be simple and easy.

They should be able to anticipate SPAM, that is to say, they should make the MTA intelligent enough to detect prospective SPAM still at the envelope command level (MAIL FROM and RCPT TO). SPAM should also be detected by real-time analysis of message headers (I plan to do it in future versions of LibAntispam), even before the message contents are received.

They should create an advanced front to maximize the protection of network output bandwidth, as well as the protection of network resources and users. This protection focuses on eliminating the impairments caused by SPAMMING techniques.

They should provide friendly tools and GUI interfaces to the network administrators and users. These tools and GUI interfaces (i.e. LibAntispam-GUI) can be used to manage, activate and deactivate ANTISPAM protections whenever they are needed. Such a management can be done either globally (applied to all users) or particularly (for each user).

They should be able, if you want, to live with others antispam soluctions that you already use. For example, if you already use spamassassim and want to continue to use it, you can do it. But LibAntispam tools and GUI interfaces cannot control spamassassim features at administrator or users level.

Paying close attention to these guidelines, it was my pleasure to create and develop LibAntispam. It will now be also my pleasure to spread it to the world.