I'm Mark Bernstein, chief scientist at Eastgate. I've wanted to build Tinderbox for years. It's exciting to see it become a reality, to have a chance to use it every day.

I've been working on hypertext tools since the early days, years before the Web. I was at the first hypertext conference, where tiny Eastgate shared a demo machine with behemoth Bell Labs. I've been program co-chair of the Hypertext Conference twice, I've written lots of hypertext research papers, and I've published some of the best known literary hypertexts. Right before starting Tinderbox, I led a complete redesign of Storyspace, ensuring that this pioneering medium for hypertext narrative would remain fresh and vital for years to come.

Tinderbox is a new tool, a new kind of software. There aren't many programs like it; that's part of the reason I find Tinderbox so exciting. It lets me do things I've always needed to do -- take notes, write memos and research reports -- in new ways. As I improve and refine my agents, Tinderbox gets better and more helpful. It's fun to use. And, thanks to its embrace of XML, Tinderbox works nicely with lots of other tools I use every day.

Tinderbox is also a deep tool. There are lots of ways to do things in Tinderbox; we're still learning just how to use it. It's like the early days of personal computers, the time people first realized that a personal computer wasn't just for games and your checkbook. It's like the early days of the Web, the time people realized the Web was good for a lot more than exchanging unpublished Physics papers. This can be frustrating at times -- Tinderbox makes you stop and think. But it's also frequently rewarding.

Tinderbox has given me a chance to harness interesting ideas from a generation of hypertext research at labs all over the world, ideas that the big Web browsers have often neglected. The biggest influence, no doubt, is Storyspace, the once-obscure system for hypertext narrative that has sometimes seemed synonymous with serious literary hypertext. The Storyspace map view has proven to be enormously useful and durable, letting writers express relationships by clustering as well as by linking. Other spatial hypertext systems -- especially VIKI and VKB (Cathy Marshall, then at Xerox PARC and Frank Shipman, Texas A&M) and ART (Kumiyo Nakakoji, NIST) provided inspiration and encouragement as well. The export template mechanism was sketched in a long discussion at Hypertext '98 with Marc and Jocelyn Nanard (Montpellier) and Daniel Schwabe (PUC, Brazil), and the Nanard's brilliant work on MacWeb demonstrated that Tinderbox agents were in fact viable. Elli Mylonas, David Durand, and Steve DeRose motivated the central role of XML. Mitch Kapor's Agenda was an early inspiration, and James Fallows demonstrated, in essays on Agenda and Zoot, that writers could and would use sophisticated agents.

Exciting software usually comes from small teams; and Team Tinderbox was so small that its members were frequently overwhelmed. Eric Cohen wrote the manual, Charlie Bennett launched the test program, Diane Greco coordinated testing and development. An extraordinary test team helped refine Tinderbox and identified a host of challenges.

Home \| News \| Download
	Eastgate Systems Inc., 134 Main Street Watertown MA 02472 USA (617) 924-9044	info@eastgate.com
© Copyright 2002 by Eastgate Systems, Inc. All Rights Reserved.