Power PDC

Trying to make sense of the recent Professional Developers Conference (PDC) held in Los Angeles is proving difficult, even after being back in the office for two weeks. My problem isn’t that Microsoft was giving out mixed or conflicting messages – far from it, the message was clear and coherent. And it wasn’t that it didn’t fit together – like some giant jigsaw puzzle, all the pieces have been unveiled and placed on the board in their proper places. No, the problem is the sheer magnitude of what it plans for the next couple of years. Trying to work out which are the key pieces, and the major thrust of many of the announcements, is hard work, especially since we know there are more pieces to be unveiled over the coming months.

Where to start? Well, for once, the keynote speakers gave quite compelling presentations, but the rate at which the unveilings followed one another made my head spin. So here’s a first attempt to grasp the technologies and news behind the PDC and some related announcements. Three things stood out: Office 12, LINQ (Language Independent Querying) and workflow.

Unfortunately, much of what’s coming our way in Office 12 is still under wraps, and that’s in addition to the astonishing changes and improvements that have already been announced. Office 12 is a huge break with the past, and it’s good to see Microsoft is finally prepared to grasp the nettle by making much-needed improvements in defiance of the sheer inertia imposed by such a huge installed user base. The new user interface components represent a significant leap forward, moving Office toward a task-based set of tools, and they’ll cause some upset among the training companies that have grown up to teach users how to use Office. Sending people away on training courses isn’t very cost effective, and today most people already know how to use the basic Windows interface. Finding even some of the more important functions in the current Office is hard, because they’re buried deep in the menu labyrinth. Office 12 surfaces the basic stuff much closer to the user, which is a big step in the right direction for those of us who believe that software should be self-teaching.

However, it’s the server-side components that will be keeping us busy next year. The ability to access Excel functionality within SharePoint, for example, is going to add a huge new twist to the Excel/web story. The increased spreadsheet size capability means we’ll be able to build much larger models than at present, and the database integration/browsing functionality is something I look forward to using as well.

I’ll be covering LINQ in more detail in a future column. This is a stunningly simple new development capability within the .NET framework, and thus available from both C# and Visual Basic, which enables a radical new way of doing data querying from your applications. It’s best thought of as a way to abstract all the messy grunge away from recordset queries, database joins and so forth. LINQ is neutral toward the format of its incoming data too, so you can join an SQL table with an XML recordset or an in-memory array. Anders Hejlsberg, father of Turbo Pascal and Delphi and now Redmond’s language guru, is the brain behind it. Not since Visual Basic 3 first arrived have I seen such a simplification in the way of dealing with data, and I think this will be a killer development tool for anything to do with data retrieval.

The third killer technology was the announcement of an OS-level workflow capability. For far too long, workflow management has been some sort of gargoyle that’s bolted onto the side of an application, with each application equipped with its own incompatible workflow engine. Building any sort of integrated workflow has become an exercise in extreme futility. However, Microsoft has announced a complete framework within the OS for managing system- and application-level workflows, which is trivially easy to use and will soon become one of those ‘yes, of course it uses that’ functions. Best of all, it removes the excuse from software vendors, as there’s no conceivable reason to not support it.