Fast or slow: Olive Software's ActivePaper allows for the display of a quick download of just text (top) or the leisurely download of the actual page (bottom).

An iota of a difference

We called it the "Holy Grail" of newspaper archiving.

In 1994, this august journal went gaga over some new technology brought forth by a small Israeli company called Iota Industries Ltd. In short, the system could search through a database of full-page images, read the text from the page and display results in the story's original format and look and feel.

Iota made a quick appearance in another suppliers' booth at NEXPO '95 and we were floored. Unfortunately, the company shifted its focus from the newspaper industry and we never heard from it again.

That is, until I set about interviewing the head of research and development for Olive Software Inc. of Norfolk, Va., an up-and-coming player in the worlds populated by companies like NewsStand and Bell and Howell.

"We've met before," said Yoni Stern when I introduced myself over the long-distance telephone. "I founded Iota."

Stern had sold Iota to an American company, and that concern shifted focus away from newspapers. A couple of years back Stern left Iota and founded Olive. "I came back to my initial, real love, the newspaper business," said Stern.

Olive is a U.S. business, which has about 25 research and development people in Israel, just outside of Tel Aviv.

And while Olive's technologies are reminiscent of Iota, Stern says they are all new: he says he has registered five international patents for Olive's artificial intelligence methodologies.

"The main thing that makes Olive so compelling is that it retains the structure and look-and-feel of the newspaper in a very compressed format," said Peter Zollman, an industry consultant who served as Olive's interim chief executive for most of 2000.

Zollman, who remains a consultant to the company, said that Olive's software can "digitize the daily newspaper in near real time, turn it around within 30 minutes to an hour after the PDF is complete, and make it available in a multitude of formats and to a multitude of platforms on a daily basis."

Olive's products work from Portable Document Format (PDF) or Tagged Image Format Files (TIFF) and segments the newspaper page into chunks -- a headline here, a byline there, stories everywhere and a picture caption somewhere else. These chunks are then encoded in eXtensible Markup Language (XML), which means that the data then has structure. And all of it is wrapped around a web browser (Microsoft Internet Explorer for Windows temporarily; others in the future).

The method of turning an image into a text file is "proprietary," said Stern. It's "a special kind of artificial-intelligence system that's trained to read a newspaper page much the way a human eye does."

In the archiving application (called ActivePaper Archive), a search for a reporter's name brings up either a "fast" representation of the story (which is transmitted in XML and uses the fonts from the personal computer) with the byline highlighted, or it can display the story as a clipping or in the context of the whole page. This product competes with the services from Bell and Howell Co. and Cold North Wind Inc.

The ActivePaper Daily is another version of the technology, designed to allow for a page-by-page display of the current day's newspaper. As opposed to the NewsStand product, which requires the complete paper to be downloaded, the Olive software brings only the pages a user desires to the web browser, which reduces bandwidth. Future features include a "my paper" function, which could conceivably download, overnight, specific topics and pages the user has chosen in advance, making the experience even faster.

Stern said that Olive was working with the Audit Bureau of Circulations of Schaumberg, Ill., to certify its system so that papers can include ActivePaper Daily readers in their circulation reports.

The ActivePaper Daily maintains a link with the reader's computer and also allows a publisher to track and analyze which stories and ads are being read. The XML structure also allows a paper to sidestep copyright issues: if the publisher doesn't have the rights for on-line distribution of some specific content, the ActivePaper product can either substitute another story in its place or merely block the viewing of the story on the page.

And while the main output of ActivePaper Daily is this web-based subscription system, once the Olive software has converted the newspaper page into XML, it can become pretty much anything -- what the Olive people call "multichannel publishing."

Both Stern and Zollman emphasized that the Olive product is just that -- a product. "We are a technology company, we sell products," said Stern. Either application can be branded so that the reader is unaware of Olive's participation.

With no commitments for ActivePaper Daily yet -- "there are active discussions underway," said Zollman -- the first rolled out product will be for the British Library and will emphasize the concept of "collections," which can be a group of periodicals or a group of concepts.

"Newspapers are more than text," said Stern. "You read what you like, not computer ASCII. I'm preaching all the time, newspapers have to go back to their original identity, which has been developed over 400 years."

Olive Software Inc., (757) 468-1903, e-mail: info@olive-soft.com.

-- dmc

From THE COLE PAPERS, April 2001, Copyright © 2001, All Rights Reserved.