By James Kobielus
This week at Microsoft’s annual BI conference in Seattle, they shared a lot of near futures from their business intelligence (BI) and data warehousing (DW) roadmaps.
In his latest post, Boris Evelson nicely described Microsoft’s splashiest near-futures demo, for ‘Project Gemini." This set of features, due for 2010 availability under SQL Server "Project Kilimanjaro," will support (brace yourselves, this gets mildly run-on) self-service, interactive, ad-hoc, in-memory, column-oriented, Excel-integrated, Sharepoint-integrated, SQL Server Analysis Services-integrated, collaborative, desktop-based modeling, creation, sharing, publishing, and governance of multidimensional analytic applications among non-technical users within the enterprise.
Whew! As you can imagine, Microsoft is still groping for a pithy way to characterize the emerging new post-OLAP paradigm that their "Project Gemini" will enable. Fortunately, their "Project Gemini" demo is a lot more user-friendly than that description would lead you to believe. It hangs together smoothly, both from a usability standpoint and as a powerful approach for making dimensional modeling pervasive among Information Workers. At the very least, it was an impressive and entertaining presentation, delivered in the context of a storyboard-style value proposition from the conference’s main stage.
The practical vision that Microsoft demonstrated corroborates Boris and my discussion of next-generation OLAP approaches in a forthcoming co-authored document. Though other BI vendors provide or are developing similar post-OLAP approaches on many levels, Microsoft has brought it all together into a powerful new synthesis that takes dimensional modeling away from the “rocket scientist” data modelers and puts it into the hands of any Excel power user. In the process, Microsoft’s approach has the potential to radically unclog IT departments’ OLAP development backlogs by providing users with do-it-yourself tooling integrated tightly into their BI deployments.
But, as I said, "Project Gemini" is still a near-futures work-in-progress, and, of course, depends on a full Microsoft BI stack. But seeing the demo several times–plus other futures that Microsoft demonstrated at this show, and then reading my other Forrester colleagues’ most recent blog posts–triggered some other thoughts regarding next-generation analytics environments. Indulge me for a moment (this is all in addition to Boris’ excellent sketch on next-generation BI).
Next generation? Shifting gears and speaking of human generations for a moment, I enjoyed Connie Moore’s discussion of user-interface peripherals ergonomics for the many Baby Boomers (self included) who are moving inexorably into AARP territory. Coincidentally, Microsoft at the BI conference demonstrated their new "Surface" technology, which can best be described as a touchscreen physical table with an immersive spatially oriented navigation paradigm (and running Windows Vista under the "surface"). On Surface, Microsoft demonstrated a Virtual Earth application that delivered fresh BI metrics tied to points on a 3-D onscreen map (projected on a 2-D physical display) of downtown Seattle.
What occurred to me while manipulating the large visual onscreen objects on Surface was that this tactile interface is suited to those of us who have always have been ham-handed, and/or those of us who’ll lose fine motor control and visual acuity as we age. No, it probably will never become the standard UI technology for pervasive BI, but these "very large screen, very large input touch surface" devices may become the only usable BI clients for many of us as we age. At some point, the microscopic UI of mobile devices–such as the iPhone apps that Ted Schadler discusses–may become unreadable and unusable for senior citizens attempting to access BI applications.
Micro-interfaces? Speaking of micro-everything, Gil Yehuda’s discussion of microblogging reminded me of an important point about "Project Gemini." Fundamentally, that Microsoft technology supports what one might call user-centric "micro-modeling" within Excel of fine-grained analytics applications (which Microsoft referred to alternately as "assets" and "artifacts"). One such analytic micro-app might simply define a report-style view of of a particular range of rows and columns in an Excel spreadsheet model. Another might create a reusable dashboard component that, after having been published to Sharepoint’s shared library, be repurposed by other users in their "Gemini" applications. In this way, micro-analytics might be composed by users into larger models (call them "cubes" if you wish). As a persistent module of corporate institutional memory, a "Gemini" micro-app posting (to Sharepoint/SQL Server Analysis Services) is akin to a micro-blog posting to Twitter. Just as you link from your blog to other people’s micro- and/or macro-posts, you’ll do the same with respect to their "Gemini" micro- and macro-models.
What do you think? Is Microsoft atomizing the OLAP cube to smithereens? Are the days of traditional MOLAP and ROLAP approaches, with their high priesthoods of expert data modelers and cube builders, truly numbered? And when will Microsoft roll out a tactile, Wii-style, immersive post-OLAP environment for bouncing analytics objects back and forth, game-style, along an n-dimensional topological hypercube surface?
OK, OK, I’m not waiting for the latter. I can’t even visualize it, not even with my thickest eyeglasses. I assume that a Ph.D. mathematician can, or maybe Stephen Hawking.
But it is fun to dream.