Giving on-line content providers and information aggregators the ability to offer their users succinct summaries of lengthy documents is the key to Inxight Summarizer software.
Inxight Summarizer is software component technology that summarizes a typical document in real time by extracting document titles, key sentences and phrases from ASCII and HTML documents. Summary size may be specified by the number of sentences desired or as a percentage of the document.
Designed for real-time, high-volume applications, it supports approximately 200 file formats, including Microsoft Word, Microsoft Excel, Lotus WordPro, Lotus 1-2-3, Corel Quattro Pro and Adobe PDF.
Everett Oliven, product marketing manager with Inxight Software in Palo Alto, Calif., said Inxight Summarizer eliminates the time users spend siphoning through information. It parses sentences into words and multiple word concepts and combines those with statistics.
According to the company, Inxight Summarizer processes HTML documents at 800MB per hour or the equivalent of 39,000 words per second.
“It goes through, and with dozens of statistical equations, says for example, ‘If you want five sentences that summarize this document, these would be the most indicative sentences of the document,'” Oliven said.
Inxight Summarizer supports Windows NT and 95 and Solaris, and is portable to other operating systems, including MacOS and Unix variants such as Solaris 86, HP-UX, AIX, FreeBSD, Irix and Linux.
Bryan Parks, director of programs for Palo Alto, Calif.-based Altavista, said this type of functionality was the reason his company decided to integrate Inxight Summarizer when developing their own product, Altavista Discovery, a PC-based product that attaches itself to a browser, allowing searches both on the Web and a personal computer.
“It provided us with the best solution for what we were looking for,” Parks said.
“We had looked at competitive products out there, but there was nothing that came close to what Inxight was able to deliver in terms of looking through an entire document, parsing out some of the more commonly used words in that document and making some kind of assumptions and some actual mathematical equations about what would be the best relevant content to appear on that particular page.”
Altavista was able to integrate Summarizer with the Discovery toolbar, Parks said, which consists of several different buttons a user can select, one of which contains the Summarizer functionality underneath. Rather than scroll through multiple lines of text, one click of the button allows the Summarizer functionality to take over.
The result, Parks explained, “is a page which is essentially a thumbnail of the entire long document in basically a couple of paragraphs.”
According to Parks, the major advantage of Summarizer’s functionality is that “we can allow customers to browse the Web much more effectively and much more efficiently. Rather than having to look at the entire document and read it, by reading even three or four paragraphs they can get a quick gist.”
Inxight Summarizer also creates dynamic query-based summaries in response to customer queries. According to the company, this option allows users to specify, for example, a term such as “Microsoft stock price” and get article summaries focusing specifically on Microsoft stock prices rather than a general summary of the documents.
Geoffrey Bock, senior consultant with the Patricia Seybold Group in Boston, said the technology used by Inxight Summarizer is unique. Although there are one or two other companies trying to do similar things, he said, “the algorithms that Inxight has…are state-of-the-art for linguistic analysis.”
A tool such as Summarizer is particularly helpful now, Bock said, as on-line content providers continually try to improve upon the activities, competition and sophistication of their offerings.
“I think [Summarizer] is pretty significant because with so much access to information and so much electronic content available with so few mouse clicks, we need to be able to analyse in a more sophisticated manner,” Bock said.
Summarizer supports a number of languages, including Dutch, English, French, German, Italian, Portuguese, Spanish and Swedish.
Pricing for Inxight Summarizer (www.inxight.com/Inxight/Products/LinguistX__174/index.html#summarizer), which includes Windows DLL and Solaris library, C. API, C++ sample application and class libraries, and one language component, was unavailable.
Inxight Software Inc., a Xerox New Enterprise Co. in Palo Alto, Calif., is at (650) 852-0290.