THUNDERSTONE NEWS

August 2002 - Archive

CONTENTS

• Thunderstone at Homeland Security Expo
• Product Focus: Thunderstone Categorizer
• Tech Corner: Vortex Variable Escapement
• Subscription/unsubscription and contacts

Thunderstone customers (and prospective customers) in the U.S. government and defense sectors are invited to stop by and meet us at the Symposium on Information Sharing & Homeland Security in Philadelphia, August 19-21. We will be at booth 217 of the Technology Pavilion. Approximately 100 other technology vendors will also be exhibiting. To be admitted to the exhibit hall without charge, mention that you are a guest of Thunderstone Software. The Technology Pavilion will be open Mon.-Tues. 10 a.m. to 4 p.m., and Wed. 10 a.m. to 3 p.m., in Franklin Hall, the Downtown Marriott, 1201 Market Street.

The Thunderstone Categorizer was originally created to support our web search engine at search.thunderstone.com, and to categorize the websites into a hierarchy. It has many other uses beyond categorizing websites, and can be trained on existing documents.

Categories can be added on an as needed basis, and the categorizer can be run either in a supervised, or unsupervised mode. eWeek labs looked at a number of categorizers, including the Thunderstone Categorizer in the July 15, 2002 issue. You can find a link to the eWeek article in the News Section of our website. For more details on the Categorizer please contact us.

When printing Vortex variables, sometimes the value is HTML-escaped, and sometimes not. HTML escapement occurs only when HTML mode is on, ie. when any of these conditions occur:

• The Vortex URL has no extension, or one of the extensions .html, .htm or .wml
• The output Content-Type is text/html, eg. via an appropriate <header> statement
• HTML mode is set directly via <vxcp htmlmode on>.

Often it's useful to avoid escaping just one variable, rather than change escapement for all variables. In that case <send $var> or <fmt "%s" $var> can be used to print the variable without escapement. The <send> function prints every value of its argument, as-is, but does not support embedded nuls (ASCII 0). The <fmt> function supports full binary values, only prints the current value of its arguments, and has a wide range of other printing options.