How should FTP-processes declare their data representation format?
One idea was that user-processes could recognize a file’s data type by the inclusion of a data type flag in the descriptor heading. Another possibility according to RFC 114, was to add attributes to the heading. The RFC fails to elaborate on the issue of header attributes any further. Another idea was to use a “set data type” transaction operation specified in RFC 294 to identify the structure (type and byte size) of data. Such information could be used to parse, interpret, reconfigure, and store data. A “change data type” control transaction code was considered for use also, for subsequent data transactions over the same connection. This “change data type” code was only necessary if the structure of the connection allowed for multiple transactions. The hosts did not have to accept the representation; however, if they could not accept or provide the specified representation they would need to inform the user. [RFC 171] How were the different host architectures to handle the differences
