Day: May 20, 2008

EMC Documentum Dynamic Delivery Service (DDS)

Presented at EMC World 2008 by:
Jeroen van Rotterdam, former CEO of X-hive, General Manager for XML Solutions
Michiel Veen, heading the engineering of XML Solutions

A new product – platform or framework for dynamic content delivery.

The Repository consists of these components:

  • Content Files
  • Attribute Tables in RDBMS
  • XML Store, pure native XML DB
  • Full-text indices
  • What about Personalized content delivery?

    Three levels:

  • Static Delivery – XML – Transform using XTS
  • xPression: Static Personalized Delivery (based on Document Sciences)
  • DDS: Dynamic Personlized Deliver (Documentum DDS)
  • The DDS is using Site Caching Services to push content into a second XML Store and that creates a second stack (parallell to Documentum Content Server) on top of that to build highly interactive applications.

    Easy to build, integrate and use. For instance interactive search forms and flexible queries

    Application
    – Travel Guides Destination, Interests, Price Range
    – Car Owners Manual Market, Model, Extras
    – Corporate Policies Geolocation, Department, Role
    – Technical Pubs Market., Equipment Type, Model Number, Task

    There is something called the Logic Engiine to generate guided content navigation or collection of data using dialogs to initate the next step in navigation.

    The Stack in DDS:

    Google Web Toolkit Client API, Web Client, Admin Web Gui, End User GUI

    Web Server
    – SCS Target
    – GWT Server (Google Web Toolkit – mycket bättre än DOJO-apps)
    – JSP Tag Library (lite enklare interface än GWT för att tex stoppa en Xquery i en java-app)

    Delivery Platform Services

    Delivery Platform API

    “Apps”
    – App config (generate XForms without any coding)
    – XProc Engine (finally getting standardized, has an XML Schema)
    – XForms (new XForms engine which executes in the client, pure Javascript component to render the forms)
    – Importer (any CMS or any source, can push data to the DDS core without any programmeing through the new Site Caching Services Target)
    – Logic Engine (tool that can consume a single XML-document to generate a guided UI)

    Persitancy Layer:
    – XML Apps
    – Stored XQueries
    – XForms Instances
    – XProc instances

    XML Store in the bottom:
    – XQuery Engine
    – XSL-T/XSL-FO
    – Indexes
    – Content

    What they are trying to do is a horisontal stack which is scalable with multiple application servers.
    The content in DDS is considered to be released or published content from the Documentum repository.

    Want to build this using XML standards.

    GUI
    – XForms
    – GWT
    – Logic Engine (wizard interfaces)

    Processing
    – DOM
    – XProc
    – Logic Engine
    – Xquery (first company with a full XQuery implementation)
    – XSLT
    – XSL-FO

    Processing flow:
    XForms Instance
    Layout.xml
    Properties

    For instance extract proprerty-file based localization of display names of attributes. The XForms object references a label which is then found in a simple text file which holds the actual label. Multiple files like this can be created for each language.

    With GWT a XQuery can be connected that drives a tree structure (as in Windows Explorer) as a web page widget. Pretty cool.

    Content Server and DDS are separate installations and allows having DDS connected to a 5.3 repository.

    It is deployed with WAR-files – Using a standard Application Server Container

    They showed a rather cool example of a Web GUI with a couple of GWT-widgets that created a interface that using drag and drop could take different XML-chunks to create a new custom document which was instantly rendered to PDF using a template. Resulted in a nicely looking PDF (and HTML) with EMC layout while the chunks was just handled as text with headlines.

    In summary a really cool way of creating personalized interactive applications. However, it isn’t always easy to make up your mind how a specific kind of content is presented and provided to the users in the best way possible. There are many interesting options to choose from using the Documentum platform and I have not been able to make up my mind on this yet.

    EMC World 2008 Day 1 Part One

    The conference has started and I was up early to go to the first seminar which was called Effective Classification – From data to information but it was mainly focused around EMCs storage products and low level classification and metadata management. More or less nothing about Documentum during the first 20 minutes so I got bored and went over to Introduction to Transactional Content Management instead. That was much better and was an overview of EMCs offerings around BPM and content management. Of course focused around the traditional examples like claims management and with integration with document capture products like Captive. However, I am very much interested in TaskSpace which provides a good streamlined interface for workflows but with inline preview of associated objects. I also had no idea that there was a solution for direct-attached scanners using a web client.

    After that I had a meeting with David LeStrat and Gideon Ansell. David is PM for the new Magellan client and Gideon works with usability issues in the Documentum product line. We met last year where we provided some info around our network visualization technologies. We presented our project and what we trying to achieve and talked about how influences from the Web 2.0 movement could be used in the Documentum platform. Personalization is important just as the personal page is to actually make the user a node in the system. The current WebTop/DAM clients offers an good interface to interact with content object but still mainly around a folder structure idea. However, there is no place to enter my personal details and my skills. Magellan will offer that and in a sense provide the first step towards a community idea. I also mentioned the need to integrate Magellan with an Instant Messaging Solution. Finally we talked a little about our aspiration to provide a GIS-oriented interface to consume objects with geocoordinates. Currently this is just a mashup based on Google Maps but we of course need a solution that works without internet access. See upcoming posts for more information about Magellan and the other new interfaces.