PIDL provides an XML syntax for document transformations useful for multicast content, such as on-line newspapers. The transformations include filtering, sorting, adding and replacing sections of a generic master document to provide personalized versions for individual users. PIDL is especially optimized for IP multicasting. In addition to the XML syntax, a DOM-like API is given for triggering the transformations.
Providing an XML syntax for generic document transformations is one of the motivations for work in the XSL working group. Requirements on XSL included issues relating to indexing, such as sorting, collating, and coalescing, as well as process-side scripting. A recent XSL working draft provides a mechanism by which a "result tree is constructed from the source tree... The result tree is separate from the source tree. The structure of the result tree can be completely different from the structure of the source tree. In constructing the result tree, the source tree can be filtered and reordered, and arbitrary structure can be added." PIDL demonstrates experience in this area and suggests requirements for future work. In addition, layering on top of XSL would increase the power of the PIDL language.
PIDL layers on top of XML. The authors indicate that future work could involve integrating PIDL with XPointer, RDF or P3P. XPointer would provide a more powerful mechanism for identifying which sections of the source document are to be altered. RDF would allow the assignment of interest scores that are necessary for filtering to be expressed as metadata. P3P would allow users to control under what conditions they give out personal data necessary for personalization.
PIDL only allows the transformation of source files that contain special <Block> constructs. Generalizing PIDL to allow source documents written in any XML schema to be personalized would allow PIDL to have a broader scope.
As the PIDL Submission notes, there are serious privacy concerns inherent in allowing subscribers to view each others' customization directives. PIDL suggests that encryption can handle a requirement for privacy. To protect PIDL users' privacy, there needs to be a way for a thin client to recognize which fragments it should attempt to decrypt and which it should not -- without exposing the identity of the intended recipient.
The PIDL Submission could be integrated with the CC/PP device profile work within the W3C Mobile Access Activity. If there is a need to multicast documents with local processing for final personalization, then sharing across devices of similar characteristics could occur as often as sharing across users with similar personalizations. A general personalization framework could encompass sufficient extensibility to handle a variety of personalization technologies.
PIDL could be expanded to address internationalization issues such as collating sequences for sorting.
The W3C Membership is invited to provide feedback to NEC and to <firstname.lastname@example.org> [an archived mail repository], and to consider the submitters' suggestion to work on future PIDL issues, including multicast security issues.
Disclaimer: Placing a Submission on a Working Group agenda does
not imply endorsement by either the W3C Staff or the participants of the
Working Group, nor does it guarantee that the Working Group will agree
to take any specific action on a Submission.