The automated document Part 1
A handful of products for creating "active documents" are ensuring that data in digital documents is accurate, complete and created without a lot of fuss.
With fast-forward innovation in information technologies, every few years experts and laymen alike need to pause and take stock of their nomenclature. Take "document," for instance. The latest merger of code and content known as "active" documents has injected a new concept into the accepted notion of digital documents--that of automated change. In the IDC research report, "Active Documents: Changing How The Enterprise Works," Joshua Duhl defines this new category of documents as "XML documents that combine highly structured XML with executable behavior [in such a way that] the code, which may be XML itself or another scripting language, is contained in or attached to the document, but executed by an external processor." As a result, he adds, "it appears that the document contains both information and the actions that are necessary to act on the information."
Most active documents are forms of one type or another, so making them active documents expedites forms processing. A simple example of an active document would be a form that, when a customer enters his customer ID in the appropriate field, automatically at run time fills out boilerplate-like the customer's address, phone and social security number for many other fields, which it pulls from a backend CRM application. It might also validate that data is in the right format in certain fields, so that if, say, the user were to enter a postal code in the wrong format, the document would prompt him to enter it in the right format. Data in some of those fields might get updated automatically over time as data in the backend system--such as inventory data in an ERP system--changes over time. It might also use data it pulls into different fields to calculate new values for those fields according to built-in rules. That saves a lot of time, especially for very long forms, and guarantees data accuracy.
Active documents can also kick off business processes that are embedded in the document. For instance, after a claims adjuster fills out the same 10 fields on any insurance claim that's an active document, it might get routed automatically to his supervisor for review and, similarly, after the supervisor fills in certain fields, it's routed to her director for approval.
Because XML documents have different structures, it's difficult to store them in relational databases, so XML repository products are used as enabling technology to index and store active documents. Those are particularly effective in deconstructing XML documents into sections that can then be reconstructed with sections of different documents to form hybrid active documents. For instance, with that strategy, a user could extract data from documents in various backend systems and reassemble it into an active document. Instead of accessing each system and, at best, cutting and pasting data for the new hybrid document, the user simply accesses it and the constituent data is assembled at run time. With that method, the structure and content of the resulting active document change based on the fragments of documents selected.
Some products specialize in changing the structure and functionality--as opposed to the data primarily--available in an active document according to the company role of the user. In a complex active document, for instance, an engineer doing automotive product design might get an image interface and access to a wireline model of the car he's working on, as well as image editing tools necessary to modify the model. By contrast, the marketing person later on in the workflow who's responsible for coming up with a campaign to promote the car might get access only to a word processing interface and the product description, as well as word processing functionality for editing the text of the description. Because the tools are pulled into the document at run time, it appears as if the various functionality is actually native to the active document.
Via such fusion of code and content, says Duhl, active documents can perform tasks as diverse as validating, retrieving and archiving data, calculating values, performing transactions, initiating workflows and, in some cases, creating an audit trail of actions taken on the document. As such, he continues, they act as an interface for multiple backend applications, to which they serve as data conduits and, thus, free their users from having to learn functionality as different as that in ERP, CRM and search systems. That saves time, but it also ensures that information is accurate and complete, as well as that a workflow process is correctly followed. The result is a document that offers more useful information and more relevant structural behavior for the knowledge worker.
Some active document products
The companies in the table on page ? offer tools for creating active documents. They range from XML repositories for archiving active documents, to forms that automatically populate fields with profile data from a backend system, to documents whose data triggers stages of workflows. Most of the offerings provide multiple types of functionality--for instance, the same product could automatically pull data into given fields from backend systems and then route the active document to the user's supervisor when certain fields have been filled in.
IxiaSoft offers an XML repository with which users can index, store and retrieve XML documents. Jean Paul Chauvet, IxiaSoft's VP, says the repository can "store any well-formed XML document without having to validate the document against the schema." So companies like Ektron embed the infrastructure in their products to gain a precise search engine for XML documents that goes beyond full-text searching. The French content management company BDOC uses it to create documents on demand by indexing fragments of content in different documents as well as the documents themselves, he continues, so that data drawn from ERP, content management and forms processing applications, for example, can be assembled at run time as one active document.
The PureEdge 8.0 Suite lets users automatically pull data from backend systems. The product, explains Paul Chen, VP of marketing, is ideal for "prepopulating a form based on a user's profile that you already know and resides in a backend system." The Northwest Multiple Listing Service, (NWMLS) benefited from that capability. It's a cooperative of 1,300 Washington State real estate brokerage firms that, as an application service provider, provides 120 forms for 15,000 real estate professionals. In listing and selling a house, brokers used to have to input lots of redundant information on the relevant forms. Now with PureEdge they enter a listing number and a secure user ID, and each form is autofilled from NWMLS databases with data about buyer and seller as well as the property being sold.
Because of FileNet's proven expertise in workflow, Chris Preston, director of business process management and forms products there, is well-qualified to comment on active documents that kick off workflows. FileNet's Forms Manager integrates with FileNet's workflow and Web content management products to push and pull data to and from them, as well as initiate stages of a business process based on where the active document is in a business process. FileNet, says Preston, is most concerned with "how you activate certain assets, typically unstructured information, based on business events [so that] instead of content just sitting in a repository, it's also tied into a process."
For instance, it might be a mortgage loan approval process. Say you were applying for a mortgage loan and during the process decided to change loan types from a 25-year adjustable rate loan to a 30-year fixed rate one. With Forms Manager, Preston says, the borrower can retrieve much of the information common to both types of loans from the database where the first loan's data is stored and apply it to the second. What's more, when certain information is filled out for the final loan, he says, its process capabilities can route it automatically through the approval stage of its workflow.
Verity's LiquidOffice provides similar capabilities. For example, Johnson Controls, which manufactures automotive systems for major automobile manufacturers like Toyota, used LiquidOffice to cut the processing time of online purchase requisition forms from eight days to three, thus shortening the cycle for approvals and receipt of goods. Its British manufacturing plant had been using a complicated paper form to make upward of 200 requisitions a month, but because of the lag in processing time employees often went around the system to obtain the parts they needed. Goods ordered without requisitions could not be tracked, and employees sometimes mistakenly used unapproved suppliers. With LiquidOffice, the company transitioned to an electronic form that did database lookups, automatic calculations and embedded online guidance on filling out the form to reduce the number of fields employees completed from 25 to seven. What's more, process functionality embedded in the form now automatically routes the requisition through a three-person approval process that generates a purchase order. When the order arrives at the plant and is checked in, the form even automatically notifies the originator with an e-mail.
Blast Radius focuses as much on automatically changing the active document's structure as its content. For instance, whereas some vendors make available personal data profiles that automatically fill in certain fields in an active document, Blast Radius offers different tools for different users interacting with an active document according to their roles in a workflow. According to Michael Ferguson, VP of product strategy, its XMetal product is an XML development environment that lets the user "describe application functionality, visual style and behavior that are associated with elements in the XML schema of the XML document so that they dynamically attach to the document at run time." He adds that the product can also actually "check with the content management system, say, to see what the workflow state of the document is, what role you are playing in the workflow, what your user rights are and deliver you a completely different user interface and [editing] tools, depending on what the business rules say are appropriate."
Documents in process
By means of the rules embedded in them, all of these software mechanisms transform conventional documents into documents in process. Different views of them at different times yield different content and even different form. XML is a key ingredient in them because, explains Duhl, "since it is text, an XML document can function as a container for content and metadata … or it can be an executable program where elements function as computer instructions and associated data." Whereas now active documents are created here and there by a menagerie of products as different as XML authoring and forms design tools, Duhl predicts that they will become more pervasive--in his words, "a disruptive technology that will affect all business applications, business processes and workflows."
That may be more apparent in advanced active documents that perform activities like generating a purchase order from a parts catalog when the user clicks on the part number, or providing interactive prompts on how to fix an engine for a mechanic using an electronic maintenance manual. In any case, active documents will extend the relevance and usefulness of documents that would otherwise have a dramatically shorter life span and, in doing so automatically, will make knowledge workers much more productive.
Selected XML-Enabled Products for Creation of Active Documents
FileNet Forms Manager
Verity LiquidOffice 3.0
PureEdge PureEdge 8.0 Suite
Blast Radius XMetaL
ixiasoft TEXTML Server 3.0
John Harney is president of ASPWatch, a consultancy focusing on market, partner and technology strategy for ASPs, e-mail firstname.lastname@example.org.