Advanced File Printing or AFP is an IBM standard for describing print data, suitable for sending to large-scale printing systems for applications such as billing. Although it is a proprietary standard, a number of vendors have solutions available which can be used to allow Alfresco to index the content of AFP files or convert these to other formats such as PDF.
TargetStream provide highly scalable software for indexing AFP files, converting to PDF and more. Written in Java. Proprietary. One mention here. Java API.
Xenos Group provides the capabilities to extract metadata, index and transform print streams such as AFP for loading into ECM systems as their original format or as PDF. Is now part of their 'legacy' software.
SnowBound provide an AFP file viewer, a batch converter for converting to other formats such as PDF and an AFP Java API. The batch converter can be invoked from a command line and can therefore be easily integrated into Alfresco as a custom transformer. Evaluation copies of the software are available. Proprietary with a base price of 3,500 USD.
TallPine provide utilities for converting AFP files into XML or PDF format and more. A number of tools are available. There is a demo version available, but of limited usefulness as it has no command line support.
Compart provides scalable, SOA-oriented solution for universal (= format independent, including AFP) document processing, batch and transactional processing in one solution: from analyzing, separating, classifying, indexing, modifying, or converting to preparing for display and output on all digital and physical channels.
IBM afpan can manipulate AFP files, though it requires to run on OS/2. there are reports that it may work on Linux through the OpenRexx wrapper. In Fedora it is packaged as oorexx. I have not investigated this though.
Papyrus has a viewer here supposedly, but the e-mail with a download link never came, despite multiple attempts.
Ricoh has a commercial solution to convert afp files to pdf for example. This one has been tested for a customer prototype since we were given a trial license. As it provides a command line tool, it is easy to integrate as a 3rd party transformer.
There is a no longer maintained Perl module here. It currently has dependencies to Parse-Binary and Encode-IBM. Your distribution or CPAN may provide them, or you can install manually.
The author of the above Perl module has rewritten the module in Haskell, which you can find on GitHub. I haven't investigated this one much, but it seems to have some basic text and PDF transformations available.
Apache FOP can apparently output to AFP. It seems there is some code to parse AFP files (see e.g. here, could be a starting point. See also AFPResources.
a Google code project has some code bits that seem to relate to AFP management, see here.