Thanksgiving no holiday for EDGAR

By Ellen Messmer
Network World, 10/16/00

As most of us enjoy the long Thanksgiving weekend, IT personnel at the Securities and Exchange Commission will be racing to complete a major upgrade of EDGAR, the nation's online document system for corporate business data.

EDGAR (Electronic Gathering, Analysis and Retrieval) provides access at www.sec.gov to an open Web repository of the quarterly earnings statements and other disclosures publicly traded companies must submit.

That's the public EDGAR, so to speak, and it gets 800,000 visits daily. There is also a more private EDGAR network used at the SEC to electronically receive and review the 350,000 filings that companies transmit each year via direct dial-up modem, private lines, VPNs or the Web.

Over Thanksgiving the SEC will be making a cutover from the IBM OS/2-based network that has been the private EDGAR for five years, converting those internal EDGAR applications to a Windows NT-based network. For the first time in its history, EDGAR will be shut down for one business day when it would otherwise be open - Friday, Nov. 24 - and won't receive or disseminate filings. The goal is for the IT staff to work through the long Thanksgiving weekend to have EDGAR open for business on Monday, Nov. 27 at 8 a.m. The transition will involve the data center in Springfield, Va., and the main SEC business office in Washington, D.C., with the two connected with high-speed private lines.

"If things don't go right, we'll revert back to the old system," says EDGAR program manager Rick Heroux. But with the $22.5 million modernization effort three years in the making, there's been plenty of advanced preparation with IT partners such as systems integrator TRW to feel confident about the Thanksgiving migration.

The SEC had little choice but to find an alternative to OS/2, since third-party support has evaporated for the operating system.

"OS/2 was reliable but people have stopped supporting it," Heroux says, adding that the SEC still uses antiquated OS/2 browsers. As part of the move to NT, the internal EDGAR applications originally written in the C programming language have been rewritten in Java with better search engines and new browsers.

Putting XML to use

Another goal at the SEC is to explore how to integrate XML into processing filings under review so that business data can be pulled by machine through XML metatags that define the meaning of document content.

Since May, Heroux has been piloting the SEC into its first use of XML-based metatags on electronic documents submitted via the Web.

As a first step, the SEC has required filings made via the Web to use downloadable software for Web forms that adds these metatags when users fill them out. The SEC provides this software for free under its EDGARLink initiative. The SEC licensed the software from PureEdge.

But the Windows-based PureEdge software, which uses a proprietary data-compression technology that requires use of the PureEdge viewers, is sort of "pre-XML," Heroux says. "But it is a good transition for us from the Standard Generalized Markup Language [SGML], which we use, to XML."

Derived from SGML, XML is still evolving as a suite of standards under the auspices of the World Wide Web Consortium, and confusion reigns in determining what constitutes true standards support. "We do plan to move to XML in the future," Heroux says. "But there aren't enough document-management products using XML. And XML as a standard just hasn't solidified yet."

The SEC is also trying to meet the demands of third-party software developers and filing agents who don't want to use the Windows-based EDGARLink to submit SEC documents. As part of this struggle to bring XML into the picture, the SEC has posted the PureEdge XFDL Forms specification at the World Wide Web Consortium site and the SEC forms templates and data fields at www.sec.gov.