It has been 10 years since IDC analysts Ed Acly, Steve Garone, Brian McDonough, Henry Morris, Jacque Sweeney and I reviewed then-current enterprise software market dynamics holistically [from the perspectives of middleware infrastructure, integration servers, enterprise and knowledge-management applications, and business intelligence] to examine "Portal Mania" for all its faults and potential in the enterprise. (Danny Rimer first used the term as far as I can find on CNET in 1998, speaking of consumer portals as opposed to enteprise portals.)
The IDC opinion in 1999 read:
a convergence will occur around an enterprise portal. However, if this convergence provides integration only at the user interface (UI) level, the result is a cheap fix-each user bears the burden of sorting out the semantic inconsistencies across all of the types of information viewed in the portal. A more lasting convergence will emerge later, built on a foundation of extended process definition and common management coupled with a context manager that ranges across the data stores (for structured and unstructured data) required in support of the process.
"Later" has arrived.
I admit that I haven't followed portal server or portal UI technology closely over the intervening 10 years, so I don't pretend that Silobreaker Enterprise Suite, which was announced Tuesday, is the first example of "later" arriving. More likely it's just the first time someone - in this case Silobreaker CEO Kristofer Mansson - has played the 10-year-old "Portal Mania" vision back to me. And showed me that it works.
It should also be no surprise to me therefore that Silobreaker founders Per Lindh, chief technology officer, and Bjrn Lndahl, chief product officer, were earning their Master's of Science degree around that time in computer science and engineering from Linkping Institute of Technology in Sweden, and doing the academic work on information retrieval that underlies Silobreaker. Afterward, they worked at Framfab, went their own ways, then got back together to continue their research and kick off the popular Silobreaker news-aggregator Web site in 2006. The idea of the Web site was to demonstrate the utility of the just-announced enterprise software product. (The product had been unofficially marketed during 2008-2009.)
Echoing the 1999 IDC portal research, Mansonn says that Silobreaker is not just about doing keyword-based queries using indexing and matching queries, but it depends on analysis and abstraction using classifications with entity extraction and merging. Silobreaker calls the process "relational analysis." The new enteprise suite is available as an on-premise product that looks and feels similar to the free Web site service but adds administration interfaces, and most importantly, the ability for the enterprise to upload its own content (from other enterprise software for example), add other content feeds, (especially subscription services such as LexisNexis, for example), and add its own taxonomy. Silobreaker Enterprise Suite (and the free service) runs on the Microsoft stack.
Closing the loop back to IDC, current-day IDC search/discovery guru Sue Feldman says:
The problem for information workers today is making sense of the information that they are drowning in... The fact that (Silobreaker uses) multiple visual tools, and linguistic analysis, as well as pattern recognition means that they will be able to surface concepts and names of people, places and things. This in turn should save workers time, and help them to discover trends and facts in the information that they didn't know were there.
Based on Sue's opinion, this product deserves a close look. Make Ed, Steve, Brian, Henry, Jacki and me look good -- if a bit premature -- 10 years later.