The 1968 Woodstock music festival was billed as three days of peace, love and music. The LegalXML Conference, which convened at the Massachusetts Institute of Technology November 16-18, 2000, could be called three days of policy, law and code. Despite the superficial differences of time, place and functioning restrooms, the two events may be close at heart in fermenting revolutionary change. An informed eye might point out that both events share a secret motivation in common: they are, at base, driven by the potential for rich profits.

The most radical idea put forth at the Conference was that expert systems, in the realm of artificial intelligence, could eventually replace lawyers by referencing codes that indicate “consideration” or the “choice of law” clause in a contract. The notion of a future with litigators locked out of dispute resolution likely pleases some, but may conjure up for others the stark images with modified dialogue from Stanley Kubrick’s movie, 2001: “Open the courthouse door HAL,” and the calm reply, “I’m sorry, Dave. I can’t do that.” Luckily, there is a host of benefits to reap in the legal industry before HAL or its silicon-based progenitors will be able decree, “first thing we do, banish all the lawyers from the courthouse.”

The purpose of the Conference at MIT was to set standards for legal documents in extensible markup language (XML), which is a sort of code similar to the Web’s hypertext markup language (HTML). The user does not see XML when reading the text of a brief or article, but it can be used to set format and, more powerfully, can allow computers to automate the movement and handling of documents that would otherwise require human mediation. The LegalXML group is seeking to create standards accepted by the market. This translates to making the XML look like the well-known HTML but with greater flexibility and power.

With recent announcements from Microsoft that it was building XML into its desktop applications, the markup language is building momentum as the best approach to improve efficiency in handling electronic text.

The Promise of XML for Courts and the Legal Industry

The significance of XML is two-fold. First, for the courts, XML would mean that large amounts of e-filing could be automated. Incoming documents from lawyers would docket themselves and, like so many well-behaved schoolchildren in the hallways between classes, follow the proper channels to the court file and assigned judge’s desktop. The busier the court, the more important it is to automate the process, so bankruptcy courts, and their vast oceans of notices and claims, have the most to gain from the technology, followed by the larger urban trial courts.

The LegalXML group is leading the effort to establish standards on document format. With over 700 members participating in the online discussions at LegalXML.org, the Conference at MIT was a chance for face-to-face conversation to iron out issues effecting a wide spectrum of legal documents, and it drew some of the leading scholars in the field. Attendees included Richard Himes, formerly of the New Mexico federal courts and now Chief Scientist with a private e-filing vendor, @Court, based in Alameda, California; Daniel Greenwood, co-director of the MIT E-Commerce Architecture Project; Donald Bergeron of Lexis-Nexis: and Jim Keane, a legal technology consultant and former general counsel with a private e-filing vendor, JusticeLink.

At the Conference on Friday, November 18, Joseph Reagle of the World Wide Web Consortium gave a presentation on their efforts to set XML standards for a broad range of industries, many of which have a lengthy lead on the legal arena in automating data transfer. The director of the LegalXML group, Todd Vincent, explained that judges in Australia are currently authoring judgments in Microsoft Word, saving them in HTML, and storing them centrally on a server in Sydney. The courts want to save the judgments in XML but they need a standard that will allow other agencies and law office software systems to read the data.

Methods of Moving Documents

The effort to pass documents between courts and other public agencies, or between lawyers and the courts involves two processes: the pipeline over which the documents travel; and the format of the documents.

The options for the pipeline are relatively limited and obvious. Public networks such as the Internet or a virtual private network (VPNs) running on the Net, have the advantage of low cost, access to a large audience and ease of transmitting documents. Local area or wide area networks offer the benefit of security, and the potential for higher bandwidth and better operational predictability (e.g., if you spend enough, a LAN or WAN can be up and running smoothly more than the Net). The drawback of LANs and WANs is a cost so extraordinarily high it would make even the Pentagon wince, so to pursue such private networks one must typically spend even more on a consulting report justifying the private network approach as “cost effective.” The final method for electronic document delivery is the so-called “sneaker net,” which consists of trading diskettes or CD-ROMs, such as the systems for filing electronic briefs in North Dakota and several circuits of the U.S. Courts of Appeal.

We have a relatively solid understanding of the pipeline for moving around legal documents. The question becomes what should we put in the pipe? In comparison, courts and lawyers know relatively little about document format. More precisely, we know that submitting smart documents offers great advantages over trading the plain text.

Past efforts to improve document trading processes are discouraging. The struggles with document format are indicated by the tiny victories that we may later look back on with regret as a loss of diversity. Examples that come to mind include the reduction over the past few years in law offices of both Corel Word Perfect software and Macintosh computers, so that finally email attachments in Microsoft Word can be sent from PCs and opened without much trouble, that is, assuming you have purchased the latest version of Office 2000. Another example is the disturbing near fetish by government agencies of clinging to Adobe Acrobat Portable Document Format-hey, Governor, I made it look like paper and, see there at the bottom, it has page numbers!-never mind that PDF turns documents into a sludge-like substance akin to electronic concrete, difficult to recycle, tweak or change. Of course, these small, essentially sideways steps came about because for several years the best way to exchange electronic legal documents was to print them on paper, fax it to the recipient, and retype them on the other end. Perhaps the greatest challenge of XML is teaching the legal industry to have hope again that computer automation can provide significant efficiency.

Elements of a Legal XML Document

In a technical workgroup session on Saturday, November 18, Todd Vincent, gave a presentation on the nature of XML documents. A document has three elements, Vincent said. There is a logical structure, data, and formatting. The fact that a pleading begins with the party names at the top of the first, page, and the plaintiff’s names preceding the defendant’s, is an example of logical structure. The data is the meat of the document, or the text that makes up the contract or legal argument. The formatting may include page numbers or line numbers along the left-hand margin.

A legal document is a combination of XML and style sheets. An XML document is data with no formatting, while style sheets provide formatting. With extensible style sheets (XSS), we are 80% concerned with formatting for output and 20% we are translating from one system to another, said John McClure, a programmer from Port Townsend, Washington. Richard Himes of @Court commented that the difference between style sheets and XML was a matter of presentation versus getting a filing into the court and populating a case management database. The latter concern, covered by XML, is far more important to the courts, Himes said. Vincent explained that style sheets and formatting are important issues in the legal industry but are outside the purview of the LegalXML group.

There are several requirements to establish XML standards for legal documents. First, it is important to have standards across legal documents, so an XML tag for contracts can be reused for a subpoena and transcript; second, if a standard is not possible, then keep the definitions as similar as possible; and, third, simplicity is essential.

Two additional elements are that XML support the legal vocabulary and, finally, XML must fit the word processor paradigm. It would prove unmanageable to retrain users to create documents in a new method, no matter the downstream efficiency, and so authoring tools for XML must look and be intuitively equivalent to word processors, Vincent said. This means XML needs to recognize tables, outlines and paragraphs. It is unrealistic to think lawyers will learn XML to mark up a document or see a Data Type Definition (DTD), so the word processor will need to apply the tags. For example, the XML-enabled word processor may recognize the format of a party name and apply the mark up without prompting, a court-mandated form could have data tags around an address field in a child support order, or a lawyer could highlight a phrase and click on a menu button to label it.

Lawyers from Australia and Germany who are developing software systems for the legal industry in their countries attended the conference. The technical language was being refined as the Conference proceeded. At one point, Murk Muller, a Dutch attorney practicing in Berlin, Germany, asked about a seeming new use of the word “semantics” in the XML context, and the audience went back and forth on the precise meaning. There are unanswered questions. As Richard Himes noted, what happens to a smart document is the XML code is accidentally misapplied, and the system then routes the pleading to the wrong docket or case file. Human review and checking will likely be required for some time to come.