The Consortium of Filing Agents and Software Developers
January 16, 2003
Mr. Jonathan G. Katz
Dear Mr. Katz:
The Consortium of Filing Agents and Software Developers appreciates the opportunity to provide our comments on the subject rule proposal.
Our interpretation of both the proposed rule as well as information presented at the open meeting held December 18, leads us to conclude that the SEC's approach to Section 16 filings on EDGAR will be such that the SEC will receive and disseminate only a data set tagged in XFDL. Third-party disseminators as well as the SEC's own public web site will be required to perform a translation or transformation of the tagged data in order to produce a user-friendly document for display and print purposes.
We fully understand the SEC's need for tagged data with respect to Section 16 filings, and we commend the SEC for seeking to take advantage of the most current trends in information processing technology. However, we must also express our collective concern about the wisdom of adopting such a radically different EDGAR filing model at this time and with forms that are of such critical importance.
Success with the approach the SEC is proposing is entirely dependent upon the completeness and stability of the adopted tagging scheme. While Section 16 forms appear to be good candidates for such an experiment, the consequences of problems encountered in reconstructing a user-friendly version of the form from tagged data could be quite serious for a number of parties including the reporting individuals, the dissemination community and ultimately for the SEC.
Our consortium also understands that in many respects, the issue we are raising is one that affects the third-party dissemination community most seriously. However as the parties responsible for the creation of a great many SEC filings, we have both substantial experience in the systems area, and a very strong interest in how those filings end up being displayed across a wide variety of web sites or printed by investors, researchers and analysts.
Consequently, we strongly recommend that the SEC modify its approach with respect to its proposed Section 16 implementation and seriously consider receiving both the tagged data set as well as a properly formatted version of Forms 3, 4 and 5. If it turns out that disseminators can successfully and consistently translate the tagged XFDL instances into displayable, printable formats, the SEC could then consider dropping the formatted version of the filing.
It would not be difficult to produce both versions of the form at the point the SEC settles upon its tagging schema. Additionally, the SEC can retain the work it has already put into its own on-line Section 16 filing solution, by adding the capability to its application to also create a user-friendly display format of the filing that is generated when the filer hits the "submit" button on the service. The SEC thereby becomes the test bed for translating its XFDL-tagged documents into a displayable format. With the lessons learned, the agency can refine its approach to the point where others can have complete faith in the efficacy of using the XFDL tagging scheme as a means of creating a user-friendly version of the filing.
The phased approach we suggest above not only gives the SEC the tagged data it needs, but ensures a reliable, less risky approach to the creation of displayable and printable documents.
Our group would also like to encourage the SEC to look at some of the more open approaches to accomplishing Style Sheet translations rather than XFDL. Using XML, for example, for the input schema would permit the SEC and disseminators to apply the more widely available XSLT (Extensible Stylesheet Language Transformation) technology to produce user-friendly displays.
One final thought: we realize the SEC may be uncomfortable with the fact that in our approach, two separate versions of the same "document" exist as part of the input stream. In many respects, the identical issue exists today with HTML. Browsers display a document derived from HTML source, and the SEC has wisely elected to not comment on whether the underlying source or the displayed version is the "legal" instance. The SEC could accept both the tagged data as well as the user-friendly version with the expectation that they are identical with respect to both numeric and narrative content.
Or, the SEC might consider resolving the issue by building a solution requiring data tags that are embedded inside a single, displayable HTML version of the filing. These would not show up on a browser, but could still be validated in EDGAR acceptance processing and used to facilitate data extractions as easily as the SEC's XFDL-tagged data set.
Again, thank you for the opportunity to respond to the rule proposal. As we have in the past, our consortium is prepared to assist the SEC with additional discussion or testing should the agency conclude it would be helpful.
Sincerely and on behalf of the Consortium of Filing Agents and Software Developers,
Richard K. Bonaparte
American Financial Printing