Document management moves to the Web

In the past few years, the World Wide Web has become a standard means of outreach for federal agencies, a multimedia bulletin board for items ranging from public service announcements to tax forms. Now, however, agencies are looking for the Web to do more than simply post documents for public consumption. To support remote and traveling workers, agencies are looking to extend their document management programs to the Web, allowing staff to access and edit documents using Web browsers.

In many organizations, the Web has become so important a conduit that the quality of Web client access to a document management system may be the determining factor in selecting a document manager. And even departments that still rely on local-area network-based document management are looking at Web extensions to allow remote access.

All of the major document management programs now offer some sort of Web access, although accessing the system via a Web browser generally means giving up some important functionality, such as advanced search tools or administrative utilities. In this comparison, we look at new versions of two well-known programs: Lotus Domino.Doc 2.0 from Lotus Development Corp. and CyberDOCS 2.5 from PC DOCS Inc. Both products have made strides in bringing a more complete level of functionality to Web access.

Domino.Doc, the winner in this comparison, offers the benefit of relatively easy setup and configuration as well as groupware features lacking in CyberDOCS, most notably rules-based workflow. If your agency already is using Notes as a groupware platform, the expanded features of Domino.Doc 2.0 leave most shops with little reason to look further for a Web document management solution. The PC DOCS solution, on the other hand, is more powerful than the Lotus solution, thanks to its customizable document profiles and reporting tools, but it is also harder to install and more expensive.

Larger departments may want to consider implementing more expensive enterprise document management systems that also have Web extensions, such as Documentum Inc.'s Enterprise Document Management System and FileNET Corp. Integrated Document Management. Both solutions are installed and customized by the vendor to suit client needs. We will take a look at both products in an upcoming issue.

Domino.Doc 2.0

When Lotus released Domino.Doc 1.0 in June 1997, it seemed almost like an afterthought. The product did provide Web access to documents in Notes databases, but it offered little control over documents beyond simple check-in/check-out. With Version 2.0, which was released in the summer, Lotus has made major strides in turning Domino.Doc into a full-fledged document management system.

Domino.Doc is an add-on to Lotus' Notes Domino server. We found Domino and Domino.Doc relatively easy to install with a single exception: Configuring Hypertext Transfer Protocol access to Domino was tricky and, because of poor documentation, delayed our installation by a couple days.

Once up and running, however, we found Domino.Doc easy to use. After the Domino Enabler program is downloaded and installed, Domino. Doc integrates well with Open Document Management API-compatible Windows word processors and spreadsheets. For example, you can save documents directly from such applications to Domino.Doc. Alternatively, you can drag and drop existing files into Domino.Doc by placing them on top of the Enabler icon on your desktop. Unfortunately, this only will work for a single file at a time. If you want to perform a batch import, you'll have to use a Notes client and go to the File/Import utility.

Domino.Doc falls a bit short of the competition in the tools it provides for creating customized document profiles. For starters, you have to go to a Notes client to design document attribute forms. And Notes doesn't offer combo boxes, radio buttons and the like to employ in forms creation. Instead, you're limited to standard data fields. And you won't find the kinds of tools offered by DocsOpen, which is PC DOCS' LAN-based document management system, for automatically generating profiles while importing files.

Domino.Doc's search tools are strong. Users of LAN and Web clients can search for simple character strings and multiple words. Users also can enter Boolean and proximity operators as well as field-

delimited searches at the command line. And users of either Notes or Web clients can specify multiple file cabinets at a time to search, although this feature is a tad awkward to use because what is searched depends upon where you are. For example, if you are in a file cabinet and you launch a search, the search will only operate on that file cabinet. Finally, you can specify whether you want to see results by relevance or by date, and you can limit the number of results returned.

Advanced searches let you specify whether you want to search for any or all of the search terms used as well as to search by creation or modification date. And any search can be saved for future use.

Once you've found the documents you're after, it's a snap to check them out. And Domino.Doc will prevent other users from checking out those documents until the documents are checked back in, thereby preventing editing conflicts. Lotus also has improved Domino.Doc's version-tracking capabilities by allowing users to differentiate between entirely new versions and drafts. By employing decimal version numbers, you can check in a document as a draft (1.1, 1.2, etc.). If you use whole numbers (2.0, 3.0, etc.), the checked-in document will be handled as a new version.

The most significant enhancement in Version 2.0 of Domino.Doc is its implementation of rules-based workflow. A new Review and Approve document type makes it easy to route documents to specified users along with comments and action requirements. You can specify deadlines for approval and have reminders sent automatically to those specified users who haven't yet given their approval. On completion of the approval process, you might specify that the document be published to the Web, returned to the originator or saved as a new version.

Customizing the workflow also is easy. In Domino.Doc, the options available for a document depend upon its document "type." For example, if you want to employ the approval feature, you create a document of the Review and Approval type. To change the options of a document type, administrators can edit the template to take advantage of Domino.Doc's "event handlers." Using rules similar to those in the new Approval type, for example, you can specify that another type of document be automatically archived after an event trigger goes off.

Many of Domino.Doc's configuration options can be accessed from a Web browser, assuming you have the appropriate rights. Administrators can create new file cabinets and set access rights to cabinets and documents. And agent-based archiving lets you set up automatic archiving of documents. You can even specify how long after retrieval from an archive a document should be automatically rearchived.

Domino.Doc's security tools are more flexible in this new version. In addition to the security offered by Domino itself— including user authentication, support for RSA Data Security Inc. encryption and the Secure Socket Layer standard— Domino.Doc lets administrators and document creators selectively control access to documents, binders and file cabinets. The new version takes security further by letting creators specify users who are allowed to edit drafts. In addition, giving a user access to a "binder"— a defined set of documents— does not necessarily mean giving the user access to all documents in the binder. You can give a group blanket access to a binder and then selectively restrict specified documents.

Although Domino.Doc hasn't brought the full functionality of Notes document management to the Web yet, it is still an attractive product for two types of organizations: those already using Notes and those looking for a solution that holds down traffic across the wide-area network. Notes' database replication minimizes WAN traffic by allowing document databases to be replicated to local servers for access.

CyberDOCS 2.5

If you're looking for an industrial-strength Web document management solution that's highly customizable, look no further than CyberDOCS.

CyberDOCS runs on the tried-and-true DocsOpen engine coupled with a relational database— either Microsoft Corp.'s SQL Server, Oracle Corp.'s Oracle7 or Sybase Inc.'s Sybase. This highly scalable solution can handle up to 100 million documents, with automatic load balancing of multiple document servers.

The only problem is that before you can use CyberDOCS, you have to install not only that software but also DocsOpen, DocsFusion, a relational database, Microsoft Internet Information Server 3.0 or later and the latest update of

ActivePages. And it's not just the number of products needing installation that is daunting. At nearly every turn, we encountered problems configuring one or another module, and the hard-copy and online documentation were of no help in resolving most of these problems.

Once properly installed and configured, however, CyberDOCS is powerful and easy to use. The Web interface is attractively designed, and PC DOCS provides the tools needed to change the user interface to suit your needs. And it's not just the user interface that's customizable. In particular, DocsOpen offers excellent controls over the contents of document profile forms. Using the program's designer tools, you can generate new forms as well as edit existing ones, employing a selection of traditional data fields, check boxes, combo boxes, radio buttons, validating fields and lookup tables. Many pre-designed forms are provided to get users started, including forms specially designed for government agencies.

CyberDOCS' search tools are exemplary. Unlike with Domino.Doc, all data fields from document profiles are searchable from Web browsers. In addition to profile field searches, CyberDOCS supports full-text searches using the Verity Inc. search engine. Users can enter Boolean arguments and proximity arguments, use wildcards and the like. Unfortunately, however, this is all done using command-line arguments.

Whatever combination of search tools you employ, CyberDOCS can sort the results by a broad number of fields. You can search across multiple libraries at once, and you can save search arguments for quick access in the future. Unfortunately, the ability to save QuickSearches and proj-ects will not be available to those using Web browsers until the next release.

Administrators will appreciate the fact that the text indexer can be run on a different server if desired. In fact, if you have an especially heavy document load, you may want to run multiple indexers for a single library. We found, however, that configuration of the indexer was hobbled by poor documentation and design. We had to call technical support simply to find out how to mark documents for indexing.

The one area of functionality where CyberDOCS falls notably short of the competition is in workflow tools.

CyberDOCS, like DocsOpen clients on the LAN, lets users check documents in and out; when a document is checked out, it is locked to other users. But there are no provisions for routing documents to other users or for creating more complex rules-based workflows, although the product can be integrated with third-party workflow solutions.

On the plus side, however, complete audit trails are provided, making it easy to see who has done what with documents.

And DocsOpen provides a very strong utility for performing batch imports of documents. The administrator simply specifies the files to import and the document profile to use. Finally, the administrator matches the fields in the profile with variables to extract from the document file. To change variables, you double-click on the variable you want to change and then either select a new variable or enter a constant value. The procedure isn't quite as easy to perform as it could be, but it is powerful.

Bear in mind that documents imported to DocsOpen/CyberDOCS are not converted to Hypertext Markup Language. Instead, users must have the appropriate application on their local machines to view the attached documents.

CyberDOCS is administered by Windows client software rather than from a browser. However, administrators will find that even if they have to go to Windows, the toolset is a very strong one. In addition to an automatic archive and restore utility, administrators will find a solid set of activity-reporting tools and a cost-recovery module.

Security is also strong. Administrators can rely on DocsOpen's password authentication or on network password authentication. The program offers group-level security for access to library and system utilities, and you can set default access rights for all users and then override them for specified groups. You also can exercise fine control over what users can do with documents, including check-in/check-out, copying, deleting and saving to remote libraries as well as access to projects and QuickSearches.

Finally, PC DOCS offers an optional Document Sentry Agent that acts as a firewall between users and documents.

Before trying to implement CyberDOCS, information systems departments will want to consider hiring a consultant with experience setting up the product. But organizations that don't need built-in workflow tools will find DocsOpen/CyberDOCS to be a powerful solution for Web document management and one that is easy for end users.

*********************************************************************

How We Tested Web Document Management Software

Document management programs include the capabilities of file indexing programs, but they go beyond simple indexing, searching and viewing tools. Document managers also provide tools for tracking revisions of files, checking files in and out, and automatically reindexing the documents when they are checked back in.

To test these programs, we created six broad performance categories: installation and configuration, document processing, search and retrieval, administration, security and workflow.

**

Installation/Configuration

Each program was installed on a network server and on individual workstations. We noted the ease of installation, looking in particular at whether the program provided information about the amount of space required and available on the appropriate drives. In addition, we assessed the ability to customize the program and its support of industry standards. Finally, we checked for cross-platform support, the variety of clients offered and the scalability of the program.

To receive a score of satisfactory, the program had to install without undue difficulty and could not require an excessive amount of system resources to run. The program also had to offer a Windows client. Programs that offered greater customizability, easy scalability, additional clients and broader network and server-platform support earned extra points.

**

Document Processing

In this category, we tested each program's tools for indexing files and for creating and modifying document "profiles" that include searchable data fields; scanning and recognizing documents; and editing files and updating the indexes.

To receive a score of satisfactory, the program had allow the user to specify multiple files for batch import. To receive a score of good, the program had to allow the administrator to customize profile fields. In addition, the program had to allow the user to specify variables for extracting DOS attributes from files for insertion into profiles during import. The program also had to provide some means of launching files for editing and then updating the indexes.

Programs earned extra points for providing notable ease of use and broader integration with applications.

**

Search/Retrieval

In this performance category, we tested each program's search tools, conducting a variety of searches, and we looked for ease-of-use features, such as catalogs for storing searches for future use. We also examined the program's tools for viewing and/or sorting lists of retrieved files.

To receive a score of satisfactory, the program had to support profile and full-text searches. In addition, it had to allow basic Boolean searches as well as simple nested Boolean searches, and it had to return a list of retrieved files that the user can scan to select a file to view. To receive a score of good, the program had to allow users to save searches in a catalog and had to allow the user to mix profile and full-text searches. The program also had to allow users to search for documents within specified date ranges.

Programs that provided additional search tools and ease-of-use features, such as argument trees, received additional points.

**

Administration

In testing administration tools, we attempted to perform backup and archiving chores, move documents to another server and call documents back from remote storage. We also examined each program's security measures and its tools for maintaining users.

To receive a score of satisfactory, the program had to provide a utility for ensuring the integrity of the indexes. The program also had to offer record locking when a document was being edited by another user. To receive a score of very good, the program had to allow administrators to perform hands-off, rule-based archiving and to back up the network drives without shutting out users.

**

Security

In this category, we looked for measures provided by the program to ensure protection against access by unauthorized users. Among the things we looked for were password protection and control over varying levels of access by users and/or groups of users. We also looked for data encryption across the Internet.

To receive a score of satisfactory, the program at least had to provide password protection for program accessing, across the LAN and across the Internet. Additional measures earned higher grades.

**

Workflow Tools

In this category, we tested the tools provided for enabling workgroups to cooperate in sharing and editing documents. First, we tested each program's ability to track multiple versions of documents as they were created and modified. Next, we looked for document-routing capabilities. We also examined each program's integration with e-mail programs, checking to see whether the program was able to track files that had been checked out and mailed. Finally, we checked each program's printing tools.

To receive a score of satisfactory, the program at least had to provide basic version tracking and access. To receive a score of good, the program had to allow users to check documents out to others via e-mail and had to retain a record of such actions. To receive a score of very good, the program had to provide document-routing tools.

**

Documentation

At a minimum, the documentation had to tell us how to set up and use the program. Comprehensive, well-organized, well-written and liberally illustrated manuals and online documentation earned higher scores.

**

Price

We assigned scores in this category according to the following costs for supporting 100 users:

Excellent: $1 to $5,000Very good: $5,001 to $10,000Good: $10,001 to $20,000Satisfactory: $20,001 to $40,000Poor: More than $40,000