• A content management repository may be an Arbortext Publishing Engine client. That is, it can ask Arbortext Publishing Engine to publish documents based upon workflow events. For instance, if you check in a new version of a document, the repository could ask Arbortext Publishing Engine to produce a PDF.

• An Arbortext PE sub-process may use an adapter to retrieve information from a repository, or to store data directory into a repository.

• The two previous approaches can work together. The repository might ask Arbortext Publishing Engine to produce a PDF and pass a POID, rather than a document. In such a case, the Arbortext PE sub-process would have to be connected to the repository to retrieve the document, publish it, and (possibly) store the result directly into the repository.