Twenty-one years ago, the first formal software recommendation that I ever wrote was an in-house memo recommending our companys adoption of a 2-year-old product (barely out of diapers) called AutoCAD. That PC-based technical drawing tool had a solid foundation of floating-point data representation for drawings over a wide range of real-world size, combined with ease of customization with its internal programming tools, but the deal-maker--as I presented it at that time--was in the ability of other applications to generate ASCII files in the format that AutoCAD could render and edit as drawings.
That was 12 years before the 1996 presentation of the first draft of whats now XML, but its been worth the wait to see this core idea of inter-application data exchange turn into the basic premise of future enterprise information management. During that transition, AutoCADs vendor hasnt always been an exemplar of openness--but wherever I look, things are moving in that direction, not only at Autodesk but also at Microsoft.
I expected, back in the 1980s, soon to see custom in-house applications performing engineering calculations and generating AutoCAD drawings as their output, or taking drawing files as input and producing bills of material and other project summaries and design analyses. Last week, though, I saw AutoCAD used in a demonstration at Microsofts Professional Developers Conference in Los Angeles that went beyond these hopes, automating not only the draftsmans duties but also the component buyers task of going out for bids on alternative materials or other customizations--and also feeding directly into visual presentations to the prospective online buyer. From design requirement, to drawing, to end-to-end supply chain integration is a path that Ive long hoped to follow with mainstream off-the-shelf systems.
In his opening keynote at PDC, Chairman and Chief Software Architect Bill Gates talked about the companys entry into its third generation of XML, progressing from data storage through in-core data representation to the fully integrated use of XML for data movement throughout both the core and the application stack. The implications for programmer productivity, and the resulting ease of delivering data in the way thats most relevant to what the user is doing and in what context, are massive. Gates made the two-pronged promise of "rich links and rich replication" as the benefits of what he called "the deep idea of schematized information": Im not an easy mark for any vendors promises, but what I saw at PDC strongly suggested that these are phrases backed up by solid engineering and careful thought about developer and end-user benefits.
I was further reminded of my 1984 AutoCAD memo when Microsoft VP Chris Capossela told the PDC audience that a virtual folder in Microsoft Vista is merely an XML file. Before he could even finish explaining the benefits, the developer audience drowned him out with applause--they immediately "got it," realizing without further hand-holding that their own applications could use sophisticated task knowledge or user-driven logic to assemble and apply arbitrarily complex criteria for grouping and presenting resources to the user. Also well-received was the notion of dragging a document or other resource icon into various regions of the screen to "paint" that system artifact, as Capossela put it, with corresponding metadata. The meta-question of metadata--whos going to take the time to make those associations?--was well-answered. As I said, not only technical functionality but practical usability were jointly addressed by the PDC presentations.
Offering PDC attendees more immediate gratification were Microsofts rollouts of its data access integration technology, LINQ, for managed-code application writers and its Expression family of illustration and Web-design tools--which radically streamlines the interaction between those who determine the look and those who define the underlying behavior of a site.
All of these technology introductions that Ive discussed here address the question of why developers should continue to favor the Microsoft platform as the place where they work and as the target for what they produce. By that definition, or for that matter by just about any definition, one would have to call this PDC a great success
Tell me what you want to see from Microsoft, and from other platform definers and tool builders, at firstname.lastname@example.org
Check out eWEEK.coms for the latest news, reviews and analysis in programming environments and developer tools.