The DOT DMS experience and NRC's exploration of three dramatically different alternatives to automating its regulatory proceedings are producing helpful guidance for other agencies as to the best strategies for operational systems. NRC's notice-and-comment rulemaking initiative appropriately uses a discussion model. Its two adjudication initiatives appropriately encompass both page-image approaches and structured tagged, character-based approaches. DOT's DMS design and early implementation is basically sound. The procedures for capturing file documents and for associated quality control are well thought out and appear to be implemented according to the design. Using page images as opposed to character-based representation as the basic format for the system also is logical, given agency needs, the nature of file documents, and technological possibilities for creating a character-based version of scanned documents. It also is appropriate to defer implementation of large scale OCR to create a character-based database until the details of electronic filing are further along. It is important that the electronic filing aspects of the DMS project be worked out in a timely fashion so that electronic filing can begin. It imposes inefficiencies both on filers and DOT to have a system that requires highly automated filers to convert their computer data to paper, followed by DOT conversion of the paper back into electronic form.
To get real benefit from the NRC and DOT initiatives, it is important to appreciate the importance of all of them and the basic choices to be made after the demonstration projects are complete. The several approaches to automating rulemaking and adjudication are not mutually exclusive. In the long run some synthesis of all the rulemaking initiatives would be adopted for notice-and-comment rulemaking, and some combination of the tagged character-based EDI initiative and the page image initiatives would be appropriate for adjudicatory proceedings.
The DMS presents special challenges for public access because it is based on page images. In this respect DMS is a proxy for any image-based system. A typical text file is about 1500 bytes, but a page image of the same document would be on the order of 40,000 to 45,000 bytes, even after the compression effectuated by the CCIT Group 4 standard. Thus, the hardware, software, and digital communications capability adequate for retrieving government information in text form well might be inadequate for retrieving and handling information from DMS. While a citizen could carry nearly 1000 pages of information stored in text formats on a single 1.4 megabyte diskette, the same diskette would hold only about 28 DMS pages stored in image form. If a remote user has a 14.4 kilobit per second modem and an ordinary voice grade telephone line the user could download the 100 page document stored in text format in about 2 minutes, but the same document would take about an hour in image formats. The file size difference has implications not only for downloading, but also for browsing. In order to browse a document, the contents of the page must be transferred to the browser's computer. If it takes 60 seconds for the screen to be updated, browsing is infeasible.
These difficulties with remote access can be mitigated in several ways. One way is to pass the page images routinely through optical character reader ("OCR") software, generating smaller text formats for each textual page. Even though current OCR software has an error rate of two-to-five percent, the redundancy in typical textual documents would make the resulting text file usable for browsing and reading purposes.
Another possibility is to make the DMS information available through an appropriately sized Internet connection, allowing end users or intermediaries to retrieve the information through whatever technologies they find most appropriate and affordable. While there is concern that ordinary citizens and their attorneys lack adequate bandwidth to obtain large volumes of page-imaged information, World Wide Web applications on the Internet regularly exchange millions of graphical images through ordinary dialup telephone lines and inexpensive modems. Software is available that translates the image formats used by DOT's DMS into image formats used by popular Internet applications.
Making all of the publicly accessible material available on an Internet server permits intermediaries and requesters to make their own choices about cost tradeoffs involved in bandwidth to obtain the information. Thus, an end user wanting only small amounts of information or having sufficient time to download larger amounts of information could make that choice for him- or herself. An entrepreneurial intermediary with a particular approach to the file-size and bandwidth issues could implement that without having to make any special arrangements with DOT and without being limited by technology choices with DOT to deal with the same problems. 303
Finally, the amount of bandwidth available to individuals for digital communications has been increasing rapidly in recent years and almost certainly will continue to increase until the natural limits of a voice grade telephone circuit is reached. A voice grade telephone circuit has the capability of handling about 64 kilobits per second. Moreover, basic rate ISDN is beginning to be deployed at prices not much greater than plain old telephone service. Basic rate ISDN with two D channels makes 128 kbps available to the desktop. Therefore, it is inappropriate to design and implement the DMS assuming that individual users of docket information will not have the technical capability to access it remotely at sufficient speeds or in sufficient quantities to be useful.
There is no reason that DOT should not follow NRC rulemaking example, and move immediately to launch a small-scale demonstration project for public access through its Internet server. DOT should begin now to conduct limited pilot testing of public access to a subset of its DMS through its existing Internet server. While the existing DOT Internet server probably is inadequate to permit public access to the full content of DMS, it is adequate to permit experience to be gained with the acceptability of performance through the Internet for page images, and to test various approaches to linking the DMS files with popular Internet applications like World Wide Web. In the pilot project, header information might not be available because making it available would impose greater cost to develop appropriate linking methods. Rather, the full page image for selected dockets and documents could be made available with a simple menu entry through Gopher and World Wide Web servers for the document. Then, any user with access to a Gopher client or World Wide Web browser capable of handling TIFF formats could access the full document. Now that the DOT headquarters building in which the Office of the Secretary is located, has a T1 link to the Internet, such a pilot project is feasible; it was less feasible when the building had only a 56 kbps link to the Department's Internet service provider: CAP Access. 304
Wider spread public access to DMS in the longer term is coupled to implementation of electronic filing. When electronic filing occurs, the burdens of OCR would be reduced, and members of the public could access essentially textual documents through lower bandwidth systems with acceptable performance.
Just because the Internet exists and has technological advantages over other kinds of wide area networks does not mean that it will be the only way of connecting computers and computer networks over wide areas. Indeed, the Internet might even fade away because people prefer closed networks for security and other reasons.
Some DOT and NRC personnel have been reluctant to consider the Internet as an attractive channel for providing public access to agency docket materials. Some of these concerns are based on perceptions of security problems arising from Internet connections. Others are based on concerns with "the size of the pipe" that is available to Internet users. Still others are based on concerns about cost. As one reviewer of the draft of this report said, "I disagree with the use of Internet in accessing legal document databases. The Internet is unreliable, overused and unsecured. I recommend the use of a database management system on a Value Added Network with a secure communications frontend."
Vendors of proprietary systems can offer agencies solutions that are tailored to the specific needs of that agency and to particular types of proceedings. Sometimes these tailored solutions are available from existing vendor software; more often they are developed under contract to an agency.
The Internet's principal weaknesses involve the absence of a proven payment system, and lower levels of security and reliability, in addition to other limitations summarized in the Technological Context part of this report.
In comparing the Internet with closed-system approaches it is appropriate to consider electronic filing functions separate from dissemination and public-access functions. The different functions have different security, reliability and payment needs, although certain features, like remote system accessibility, may be desirable for both types of function.
Open systems such as the Internet offer more advantages with respect to dissemination and filing of comments in rulemaking proceedings than with respect to filing in adjudication proceedings. Security concerns are reduced, though not eliminated in dissemination systems. Signature and related authentication problems are eliminated because the identity of a member of the public accessing information does not matter. Usually, the absence of payment systems does not matter, because most agencies do not seek to collect money for access to their rulemaking and adjudicatory dockets unless the agency incurs direct costs for copying. On the other hand, any kind of open remote electronic access increases the risk of intrusion into agency systems, and this risk is certainly present with the Internet. When open-system remote access is involved, it is essential to use appropriate "firewall" measures to protect both non-public data and the record-copies of public data from alteration or tampering.
Assuming this is done, the Internet is more attractive than other obvious alternatives for allowing remote public access to agency materials of all kinds, including rulemaking and adjudication dockets. It is non- proprietary, it permits value added features to be added by different entities in public and private sectors; it facilitates competition among vendors. The competitive character of the market facilitates low-cost citizen connections with whatever hardware and software a citizen possesses. Internet-based access relieves the agency of the obligation for providing consumer-oriented telecommunications capability. In other words the agency need not worry about providing adequate dialup telephone lines, or 800 numbers, (although it must pay for an Internet connection.) It also gets the agency off the hook somewhat with respect to the "size of the pipe," notwithstanding that this is a concern of some doubters of Internet approaches. As long as the agency has a reasonably sized connection to the Internet, use of the Internet as a means of public dissemination allows users to decide on their own pipe size. Users that need high bandwidth, for example a redisseminator, can provide for it by obtaining T1 or better access to an Internet service provider. Users who want only occasional access to electronic dockets can get by with lower cost connections to the Internet. Thus, decisions about pipe size are decentralized and democratized.
With respect to electronic filings in notice-and-comment rulemaking, the Internet also is attractive, compared to closed-system alternatives because it provides a universally available, standard and non- proprietary way for lawyers and other members of the public to obtain information on agency proposals and to submit comments and documents. As with public access, concerns with signatures and authentication are reduced because anyone is entitled to submit a comment and have it considered regardless of the commenter's identity. Use of the Internet by the NRC in its rulemaking initiative almost certainly will result in wider public participation at lower costs than more traditional value added network approaches.
When adjudication is involved, the preferences may change. The specialized formats developed to automate the handling of adjudicatory submissions, as under the NRC's EDI initiatives or under the Utah court system SGML approach, can be sent through open systems like the Internet, but they fit less comfortably with Internet applications such as World Wide Web, which are oriented toward free text and images rather than toward processing structured information.
Moreover, signature and other authentication concerns are greater in the adjudication context. Enforcement of prohibitions against false statements depend on being able to prove authorship. Until the more sophisticated digital signature techniques determined in the FDA signature proposal to be appropriate for open systems are more widely available in the Internet, agencies may be driven to closed systems for operational electronic adjudication systems.
When an agency takes the closed-system approach, it must decide the degree to which the contractor developed software will be available to competitors and to the public generally as opposed to allowing the contractor to retain proprietary rights in it. ABA recommendation 109C encourages agencies to place contractor developed software in the public domain. 305
Even when contractors are allowed to retain certain proprietary rights in software developed for automated docket management systems, agencies should ensure by appropriate contract language that interfaces will be available without proprietary claims so that members of the public and competitors can obtain access to all relevant parts of the value added bundle of information made available through the contractor system.
Some rulemaking and adjudicatory notices have requested filers to limit the formats of paper filings to make them suitable for electronic filing systems. 308 The FDA is reviewing comments on its proposed rule for electronic signatures to facilitate electronic filings. 309 Many agencies have found that electronic filing greatly reduces the burden of paper forms. 310 Other agencies plan electronic filing programs. 311
Virtually all of the electronic filing requirements actually in place and contemplated have some kind of limited exception, allowing filers to avoid the electronic or filing requirement on specific showings of hardship. More generally, OMB Circular A- 130, para. 8(a)(3), 313 notes that information technology can allow agencies to meet program objectives more efficiently and effectively, specifically mentioning EDI filing for tariffs, custom declarations, license applications, tax information, and environmental reports. The circular includes, in the guidelines for consideration of electronic filing requirements, the likelihood of ready access to necessary information technology in maintenance of information in electronic form by members of the affected public, and the likelihood that mandatory electronic reporting will not impose substantial costs or other adverse affects on the public. 314
Conceptually, the permissibility of an agency filing format requirement is tested based on its tendency to eliminate the opportunity to file with respect to those entitled to file. Thus, if an agency were to require that filings be made only in the Chinese language, the fact that relatively few members of the filing community have access to Chinese would practically eliminate the opportunity to file and thus make the requirement questionable. On the other hand, requiring electronic filings by developers of personal computer word processing software would unquestionably be permissible because such entities almost certainly have access to word processing technology. 315
Any filing format requirement, including requirements that paper filings adhere to certain size and legibility requirement, and that electronic filings be accomplished by means of certain technology applications, will be judged based on a balance of the benefits to the agency and public of the requirement, weighed against the burdens imposed on persons likely to file. As in other instances of balancing formulae, the availability of less restrictive alternatives enters into the assessment. Thus, if an agency were to require filing only by means of a proprietary electronic bulletin board system, the availability of filing methods using nonproprietary Internet protocols would be one factor arguing against the appropriateness of the agency requirement.
When most members of the community expected to file have ready access to the required technology, and the only question is the availability of the technology and burdens likely to be experienced by infrequent filers, the matter can be handled by providing for exceptions to the filing requirement. Such exceptions should be available only on request and upon explanation of the reasons why compliance with the filing format requirement would be onerous, although the showing itself need not be particularly demanding. Comments solicited by trade associations and public interest groups sent on postcards or tearouts from newsletters present a special problem. Nevertheless, the repetitive nature of the content of such comments should make them relatively easy to keyboard, and associations regularly interacting with agencies can be encouraged to steer their constituents to electronic formats through email or otherwise. Another problem with a mandatory electronic filing requirement is the phenomenon of comments that are sent to another agency, to the President or to a member of Congress and forwarded to an agency.
There are many reasons why an agency might prefer an image-based automated docket management system, as the DOT has. Image-based systems present fewer risks with respect to errors generated in the acquisition stage, but they require more storage and present greater challenges for arranging public access remotely. Because image-based systems acquire information in exactly the form it is filed, they avoid the challenges and difficulties associated with specifying and enforcing document format standards through electronic data interchange ("EDI") or otherwise. They enable presentation of information exactly as it was generated by submitters, but disable full-text retrieval approaches unless additional information about documents is keyboarded or otherwise generated at the time of acquisition. Image systems, however, do not lend themselves well to electronic filing. For example, there is no way that a computer system can determine that a document was filed by ABC company from the bitmap image of the document itself; such a determination and retrieval of the document by this characteristic is possible only if the characters "ABC Co." have been separately keyboarded. 316 Conversely, a character-based system can retrieve all documents that contain the characters "ABC Co." simply from the basic representation of the document's contents. 317
When image-based systems are used, the agency faces greater challenges in designing an electronic public access system, through the Internet or otherwise. Remote public access is more challenging because the larger size of electronically stored documents makes it more difficult for members of the public to retrieve large documents or very many documents except over high bandwidth connections, which are not yet generally available. The agency must provide for electronic filing and for generation of character-based versions by OCR or otherwise so that agency personnel and members of the public can use the power of free text searching to retrieve material of interest. In addition, the particular standards used for image representation must be such as to permit the material to be retrieved over a period of many years or decades under guidelines developed by NARA.
Character-based systems require making provision for data entry unless electronic filing is used. In addition to accommodating electronic filing, they may cost less than image-based systems because they may involve less data entry by agency personnel then by image-based systems, which necessitate entering some data about filed documents.
On the other hand, any electronic filing system that hopes to provide more than basic ASCII text on agency retrieval and presentation systems and to members of the public must confront the input-data- format challenge and come up with some acceptable method of tagging and coding received documents. The approaches enjoying the most support in this regard are EDI and SGML.
A transaction set for motions to intervene regarding standing
An automated fact pleading requirement
Summary judgment motion template
Pretrial order template
A template for adjudicatory case management: with issue- oriented "slots" for all pleadings and evidence
The last of these is considerably more ambitious than the others, because it involves a higher level of generalization among diverse proceedings.
As more agency materials become available publicly in electronic form, especially as they become available through publicly accessible networks, it is appropriate to develop a new citation system that accommodates the technology. Such a citation system should be non-proprietary, thus enabling competing vendors to use it equally. It should identify the place or address from which one can obtain the information, as well as identifying a particular piece of information unambiguously.
While the details of such a system have not yet been worked out in any form of consensus among affected interests, it is reasonable to suppose that it would have the following basic elements:
It would be based on Internet universal resource locators (URLs), as are used in the present World Wide Web system.
It would identify particular parts of a document with paragraph numbers.
It would have a docket number assigned by the agency, either an existing docket number or, preferably, a new sequence number.
It would be included in the "bluebook," a publication of a consortium of law school law reviews almost universally used in the legal profession as the guide for citation form.
303 See generally Henry H. Perritt, Jr., PUBLIC INFORMATION IN THE NATIONAL INFORMATION INFRASTRUCTURE, Report to the Regulatory Information Service Center, General Services Administration, and to the Administrator of the Office of Information and Regulatory Affairs, Office of Management and Budget (May 20, 1994) (explaining how Internet can be both a wholesale and retail market for agency information when the basic information content is made available on Internet servers maintained by agencies).
304 Such an Internet public-access project also would further the Administration's NPR and NII initiatives, as the REGNET laboratory attempts to do.
305 "3. Agencies should ensure that electronic information developed with public funds, including data and retrieval software, is available to the public. When agencies contract with private entities to develop new electronic information products, the agencies should negotiate rights in data terms that will put both the data and the appropriate retrieval software to obtain that data in the public domain." ABA Recommendation No. 109C, adopted by the House of Delegates of the American Bar Association, August 12-13 (1991) (guidelines for federal and state agency dissemination of public information in electronic form). Both ABA recommendations are available in full text for viewing or downloading from the World Wide Web at http://www.vcilp.org/Aba/adminlaw.html.
306 The word "filings" in this section is intended to include all types of submissions, including petitions, motions, comments, reply comments, and requests for exception.
307 See 18 C.F.R. § 154.1(b)(c) (1994) (requiring tariff changes to be filed on an electronic medium); 18 C.F.R. § 154.26 (requiring Natural Gas Act rate schedules and tariff filings and contracts to be filed electronically, including filing by EDI); 18 C.F.R. § 157.6 (requiring applications under Natural Gas Act for certificates of public convenience and necessity and for abandonment to be filed electronically); 18 C.F.R. § 157.14 (requiring certain types of exhibits to be filed electronically); 18 C.F.R. § 385.2011 (procedures for filing electronically, applicable to various specific filings); 9 C.F.R. § 205.204 (contemplating electronic filing of effective financing statement under packers and stockyards administration rules); 17 C.F.R. § 202.3 (requiring electronic filing before SEC); 17 C.F.R. § 202.7 (same); 17 C.F.R. Part 232 - Regulation S-T (general rules and regulation for electronic filings); 19 C.F.R. § 143.31 (custom service general rules for electronic filing); 46 C.F.R. § 514.8 (Federal Maritime Commission electronic filing); 59 Fed. Reg. 64646 (Dec. 15, 1994) (ICC proposal to continue electronic filing requirements); 59 Fed. Reg. 63743 (Dec. 9, 1994) (FCC proposed rules for electronic filing of certain multipoint distribution service and instructional television filings); 59 Fed. Reg. 62705 (Dec. 6, 1994) (proposed ICC rule for electronic filing of security bonds, trust fund agreements, and insurance certificates).
308 See 59 Fed. Reg. 58810 (Nov. 15, 1994) (public meeting notice for Coast Guard proposed rulemaking for tank level or pressure monitoring device filings requesting that filings be limited to unbound format, no larger than 8 1/2 by 11 inches "suitable for copying and electronic filing); 59 Fed. Reg. 52590 (Oct. 18, 1994) (same; Coast Guard rule for inflatable life rafts); 59 Fed. Reg. 48362 (Sept. 20, 1994) (I.R.S. notices of intent to standardize individual tax returns filed on paper to ensure compatibility with scanning and machine reading of forms).
309 See 59 Fed. Reg. 45160 (Aug. 31, 1994) (proposed rule on electronic signatures and electronic records).
310 See 59 Fed. Reg. 29292 (June 6, 1994) (SEC report to OMB on effect of electronic filing).
311 See, e.g., 59 Fed. Reg. 20220 (Apr. 25, 1994) (patent and trademark office item no. 801 on regulatory agenda, contemplating electronic filing system for assignments); 59 Fed. Reg. 18188 (Apr. 15, 1994) (Social Security Administration "reengineering" of disability claims process, including electronic filing); 59 Fed. Reg. 11604 (Mar. 11, 1994) (FCC proposal to establish advisory committee on developing an electronic filing system for Common Carrier Bureau); 59 Fed. Reg. 3099 (Jan. 20, 1994) (FCC order with respect to implementing electronic filing of applications for private land mobile radio services); 59 Fed. Reg. 1455, 1456 (Jan. 11, 1994) (immigration and naturalization service expression of interest in electronic filing of application for immigration benefits).
313 58 Fed. Reg. 36068 (July 2, 1993).
314 A-130 § 8(a)(3)(b) & (f).
315 At least, this would be so if the requirement were merely that textual materials be filed on diskette in the format of word processing applications of the filer's choosing.
316 As explained infra, imaged documents usually are linked to databases with character-based representations of the document header, enabling one to find the document based on its description.
317 In either case, indexes are used to improve performance of the retrieval system.
318 See West Publishing Co. v. Mead Data Central, Inc., 799 F.2d 1219 (8th Cir. 1986) (page numbers in national reporter system reflected particular arrangement of legal materials and thus was entitled to copyright).
319 DOT Request for Records Disposition Authority, Job No. N1-398-95-3, approved by Archivist June 28, 1995.
Back to Table of Contents