When I first wrote about Tableau version 1.0, I noted this business intelligence (BI) product's twin revolutions: It works and acts like a broadly distributable desktop client, and the software is sold by the desktop, making it as straightforward to deploy as any other desktop software. The actual client's feature set was middleweight for the BI world, but quite usable, especially for end-users who have mastered When I first wrote about Tableau version 1.0, I noted this business intelligence (BI) product’s twin revolutions: It works and acts like a broadly distributable desktop client, and the software is sold by the desktop, making it as straightforward to deploy as any other desktop software. The actual client’s feature set was middleweight for the BI world, but quite usable, especially for end-users who have mastered the accursed pivot table.Tableau is now at version 2.1 and loaded with additional functions, better scalability potential, tighter work with back-end databases and a boatload of new analytical functions. I have seen a couple of demos, one guided by the Tableau Software team, and I’m very impressed with where they’ve taken the product — from middleweight to light-heavyweight in the feature department. What I don’t yet know, and what I won’t until I deploy an actual application, is whether it conforms to or bends Angus’s 2nd Law of Software: As the number of features added goes up in an arithmetic progression, the friction of using it goes up in a geometric progression. Some of the interfaces on the new features are perfect. “Quick filters,” a method for filtering hits interactively on the fly, is intuitive and snappy, as is the new feature that lets you zoom in on sections of graphical charts. The screenshot below shows results in the basic pivot-table-like interface. The shelves to the left and on top are interactive drag and drop areas. View larger imageAlso, Tableau users can now join multiple tables or other data sources on the fly, leading to easier interaction that results in more questions being tried out and possibly answered. To enhance scalability in general, Tableau 2.1 supports easy sampling of massive data sources, so the client can get a manageable slug of data that the analyst’s workstation can chew through and present faster. Added Web integration features allow Tableau work areas to reach out to the Web and grab search results and data; the results look reasonable, though I don’t think they will serve more HTTP-oriented shops better than the majority. Perhaps the most important new feature is the built-in ability to create chained analytical windows into workflows. I’m excited to try out this feature not only for the possible modular development of analysis routines but also to build structured, documented analyses for those with less ability to bend software to their cognates. Technology Industry