As part of the Linked Data Planet conference, on June 18th I'm hosting a panel described on the conference program like this:
T8: Linked Data Workshop
Based on questions submitted by the audience, a panel of Linked Data experts discuss architecture and system development issues surrounding Linked Data application development.
MODERATOR: Bob DuCharme, Solutions Architect and Author / Conference Chair, Innodata Isogen
PANELIST: Dr. Melliyal Annamalai, Principal Product Manager, Oracle
PANELIST: Michael Bergman, CEO, Zitgist LLC
PANELIST: Stefanos Damianakis, President & CEO, Netrics
PANELIST: Uche Ogbuji, Partner, Zepheira
PANELIST: Nikita Ogievetsky, Vice President, Morgan Stanley
PANELIST: Walter Perry, Managing Director, Fiduciary Automation
PANELIST: Dr. Andy Seaborne, Research Scientist, Hewlett-Packard Research Laboratories
What questions do you have? What system or architecture issues would you like to see our distinguished panel discuss? For example, if you have an idea for an application that links publicly available data using SPARQL end points or related technology, and you've worked out part of your app but are unsure about the rest, tell us what difficult parts remain. Perhaps you have a prototype that works fine, but you're wondering about the best way to scale it up. Perhaps you have an idea that's part crazy and part brilliant, and you're wondering how best to nudge it toward "useful". With this panel's well-known representatives from the tool side, the app user side, and the integrator side, it's guaranteed to have some good advice for you.
The best questions will get a full free pass to the conference, and I'll announce the winner's names here and when I read their questions at the panel. Send me your ideas for questions at bob@snee.com. (I have a few of my own ideas for questions, but I already get in for free.) There are no limits to the number of questions one person can send, but you get extra points for concision—I don't want to spend a lot of the session time reading your question out loud.
This is a fairly broad question, but: When it comes to exposing enterprise data from proprietary and COTS systems in a linkable way, two obvious choices come to mind: Provide something like a SPARQL front-end on front of each application, or aggregating all of that into a queriable "semantic data warehouse". What are the benefits and pitfalls of those approaches, and are there approaches that lie somewhere between that have proven to be successful?