The queue to e-mail symbiont has logic to extract addressing information from word processing documents printed as PostScript. These documents can either be printed locally or from network sources such as Pathworks. The first page of a document should contain addressing channel commands followed by a hard page break. The addressing information will be extracted from the PostScript in an application independent fashion and used to send the document along as e-mail. This allows users to print a document from their Macintosh, PC, etc., and have the document, for instance, sent as a FAX via PMDF-FAX. The page containing addressing information will not appear in the FAXed document.

The logic used by the symbiont is as follows:

• IF a file printed to the symbiont lacks To: addressing information THEN
  • IF the file is a PostScript file (i.e., begins with "%!") THEN
    1. Addressing information is extracted from the first page of the document, and
    2. the message is converted to a MIME multipart/mixed message and passed on to the addressing channel. The converted message has the structure shown below:

       ... RFC822 message headers ... Content-type: MULTIPART/MIXED; BOUNDARY="boundary" --boundary Content-type: TEXT/PLAIN; CHARSET=charset ... extracted addressing information ... --boundary Content-type: APPLICATION/POSTSCRIPT ... special PostScript to suppress the display of the first document page ... --boundary Content-type: APPLICATION/RMS Content-transfer-encoding: BASE64 ... PostScript document; encoded to prevent any possible damage e.g., line wrapping) while transferring through the mail system ... --boundary--

In 1. above, if the PostScript file does not follow the PostScript Document Structuring Conventions (DSC) then a second pass is made during which all PostScript strings are extracted. (In the first pass, the DSC is used to attempt to identify the beginning and end of the first page of the document.)

The character set information will be as specified with the CHARSET option. If no such option was specified, then the character set used by the local channel will be assumed.