Trusty local Syntactica software re-emerges

A local outfit has released a terrific iReader browser plug-in (Firefox and IE versions are available) that lets you hover over hyperlinks to get surprisingly useful text summaries of what lies beyond the click.

This exploits a clever Syntactica text-analysis technology that has been around for a while in various incarnations (while changing hands several times).

In fact, I wrote about the technology's creator, Arnold Schultz, back in 1996 -- and his Syntactica software brainchild goes back even further.

See my May 13, 1996 column about Schultz (who has recently had an active hand in Syntactica and iReader development) after the jump.

Update: Shut down by Google, they're baa--aack.

Update: The Possibilities of iReader by Syntactica (and its underlying engine)

ICONOVEX CONSULTANT WAS UNSUNG LANGUAGE PIONEER

Published on 05/13/1996
Section: Tech
Page: 3C
Byline: Julio Ojeda-Zapata, Staff Columnist

Arnold Schultz is getting on in years. His right hand trembles slightly as it holds a coffee cup, and his soft voice is drowned out by the loud laughter of youthful, hard-charging software developers in an adjoining room.

Schultz teeters on the brink of irrelevance. He no longer has rights to his brainchild, a software system that understands English well enough to assemble summaries and ``back-of-the-book'' indexes with ever-increasing accuracy and speed.

The Syntactica ``engine'' that Schultz pioneered a quarter-century ago is at the heart of Iconovex Corp. products, including the new WebAnchor program for indexing World Wide Web documents. But he was not among those who showcased the software at the recent Spring Internet World trade show in San Jose, Calif. 

Schultz, now a part-time consultant for the Bloomington firm along with computer expert and longtime partner Donald Malcolm, jokes that ``the only reason I'm still around is to tell people what nouns and verbs are. That's my claim to fame now.''

Iconovex workers seemed puzzled (even alarmed) when I insisted on interviewing Schultz instead of company bigwigs.

``Arnie is one of the most tenacious people I've ever met,'' says Iconovex general manager Bernardo Sotomayor. ``He's an asset to us, no doubt about that. ... But he no longer controls the software and handles no more than 30 percent of it now.''

I had initially been inclined to ignore Schultz. I rolled my eyes when he called me several weeks ago and, later, mailed me a packet of press clippings that alluded to his work. There were exclamation marks in the margins, and notations such as ``That's me ... in 1971-72,'' struck me as pathetic.

But I came around after seeing the attention Iconovex garnered at Internet World. When the company drew plaudits with its boast about ``software that understands ideas,'' I realized they were making an oblique reference to Malcolm and Schultz.

The incredible thing is, Arnie had little knowledge of linguistics or computer programming when he began work on Syntactica at the University of Minnesota. He had majored in English literature and minored in philosophy, leaving him largely ignorant of English grammar and syntax. (He had picked up the language intuitively, as had I).

When the university's anatomy department recruited him to attempt to electronically index medical abstracts - apparently because ``they thought someone who majored in English literature must be a linguist,'' Schultz recalled - he was initially forced to rely on a high-school grammar text.

He was in good company - the mainframe computers of the late 1960s couldn't make sense of English syntax, either.

``Consider, for example, the following sentence: `When anti-beef-insulin serum (guinea pig) was injected into rats, it produced a transitory rise in the blood sugar level,''' a university report of the time said.

``Unfortunately, the component words by themselves ... give little indication of the type of study carried out,'' it says. ``This study does not deal with the blood sugar level in guinea pigs. ... Much meaning is lost when the information retrieval search program is limited to the mere presence of the words themselves. Only through syntax do (their) precise relationships assume their full significance.''

Such syntactical puzzles continue to keep Iconovex experts busy today, but Schultz got the ball rolling in 1968. He wrote in his own report about his now primitive ``syntactical processor,'' which ``provides another step towards the ultimate goal of mechanization of information handling procedures.''

Eventually, Schultz and Malcolm formed their own company, Syntactic Analyzer Inc., and tried to talk corporations such as IBM and Xerox into contributing to that vision. They got praise but little money - possibly because the two men were reluctant to divulge the methods behind their magic.

Schultz recalls that an IBM executive exclaimed, ``It's a miracle!'' when the Syntactica system processed an intricate text document. But a letter from the corporation stressed that ``our communication does not in any way constitute any kind of business relation between you and IBM.''

Which brings us to the early 1990s. Syntactic Analyzer had survived thanks to dribs and drabs of venture capital, write-ups in magazines such as Business Week and the strength of its ideas, including the transfer of the Syntactica engine from bulky mainframes and minicomputers to laptops.

About two years ago, a Minnesota computer-component manufacturer sensed a future in software development and bought the rights to the Syntactica system through an intermediary. Sotomayor was brought in to manage Innovex Inc.'s new subsidiary.

Now that subsidiary, Iconovex, is on the cutting edge of software development.

On the heels of its Indexicon product for text indexing and its AnchorPage and WebAnchor programs, which do the same for Web documents, the company is readying a new product that will be a syntactic companion to Internet search engines. Raw word-search ``hits'' will be magically converted into more manageable and meaningful data, Sotomayor predicts.

Schultz expresses amazement at an online revolution that never figured into his plans for the Syntactica engine. ``I didn't know the Internet existed,'' he admits. ``I just liked playing with words.''

He is clearly worried about being forgotten amid the hoopla surrounding cyberspace, even though he gamely remarks that ``as long as they give me some money to pay my rent, I'll be happy.''

During a recent chat in an Iconovex conference room, where sleek, black, high-backed chairs make the long-haired, sweater-clad visionary seem out of place, Schultz taps the front of a shrink-wrapped Indexicon box.

``Coming from a literary background, I know how nice it is to write a book and have your name on (it),'' he says. ``But in this modern age, you write a program and don't see your name anywhere. Inventing is a very anonymous activity.''