European Semantic Desktop Project Gains Momentum

By Jennifer Zaino  March 27, 2009

Sauermann compares Gnowsis’ business model to “the typical Californian startup approach: Start small, build something useful, bring it quickly to customers, improve from there.” Think of its initial release as a "tagging platform with a little semantics” that will help Gnowsis bootstrap the market and learn how to deliver a product which it can improve incrementally with more features as time goes on. The semantic desktop will be available as a low-priced option for consumers and a higher priced system for the enterprise.

“The platform has to be both web and desktop. We are good on the desktop and have a lot of experience there, but without a good web counterpart the software does not provide everything our customers want,” he says. “When people use the semantic desktop, they soon realize that sharing their data saves a lot of time. We are currently working on a web platform that allows just this: social exchange of semantic data.”

Gnowsis plans a software-as-a-service option and free downloadable and web-based try-out versions, as well.

To turn this into a viable business, Gnowsis will first address niche markets where it can provide a good feature on top of existing solutions. “I can't say much about the exact niche market we will address first, but tagging e-mails will surely be part of it,” he says.

The founders are currently working on solving issues such as speed and memory consumption problems with the Windows and Mac versions. (A KDE Linux desktop implementation of Nepomuk's core concepts exists, but Sauermann says that most companies want a semantic desktop for Windows.) Other technical challenges revolve around building a scalable Semantic web stack for the social exchange part of the product to work.

“That is easier today using highly scalable RDF stores, but we miss a lot of tooling on top. Look at how quickly an application can be built using Ruby-On-Rails and MySQL. I think this kind of productivity is missing for RDF,” Sauermann says. “If these issues are unsolvable, there is a good workaround: just don't use RDF everywhere, fall back to relational databases for some parts.”

Sauermann says Gnowsis has been able to benefit from DFKI’s good track record of founding companies, and from the experience of many successful startups – that helped hm learn that it’s smart to build something good on something that works – and also that “building an architecture” as a means to an end can bite back. “We often end up building an architecture as a mean to satisfy our own inner nerd, so the challenge is to reduce the many possibilities to what is really useful,” he says.

He thinks Gnowsis is on the right track: “The good thing about our product is that it really addresses an itch to scratch -- remembering how the e-mails and files are connected with each other.”