Open for business

COMMERCIAL VENDORS’ embrace of open technologies is changing the IT landscape in ways no one could have imagined. Version 11 of Microsoft Office not only will read and save documents as exchangeable, human-readable XML but will validate and transform documents without special knowledge or effort. Apple is abandoning its historical preference for proprietary software and is moving to OS X, a commercial operating system built largely from open-source parts and published by Apple as open source. And then there is the open-source movement itself. Free software picked up so much steam during the recession that it can be credited with changing the direction of commercial development not only at Apple and Microsoft but at IBM and Sun as well.

Microsoft Office 11

The disruptive influence of the next major release of Office may not be felt by many of its users. People accustomed to using Office as a basic productivity suite may not care how their documents are encoded, stored, and retrieved. In many business, academic, and technical settings, however, Office is seen as a set of extensions to Windows. It has become a platform for software developers who create vertical and in-house applications. Office brought scripting and OLE (the precursor to COM) to desktop PCs. Office 11’s XML features are similarly advanced stuff. Then again, if something as advanced as the ability to drop a spreadsheet graph into a word-processor document is now taken for granted, perhaps Microsoft can push XML into the mainstream, too.

Office 11 isn’t a released product yet. Microsoft gave Test Center analysts a look at beta code. Jon Udell had the rare privilege of taking a guided tour around the new XML capabilities (see “XML for the rest of us,” Nov. 18, 2002, page 18 ). Udell is not easily impressed, and some of us expected little more than lip service to open document standards. After all, Office is the software industry’s poster child for vendor lock-in, and Microsoft managed it by keeping the details of the Word and Excel file formats secret. Programs that import or export Office files, such as Sun’s StarOffice, do so imperfectly at best. The simplest, most reliable way to get data out of an Office document is to fire up Word or Excel, a solution that suited Microsoft just fine up through Office XP.

It looks as if Microsoft will live up to its promise to add proper support for XML in Office 11. Beyond the basic “Open as XML” and “Save as XML” menu options, Office 11 will support the use of XML Schema for formatting and validation. Office 11 can enforce the structure set forth in an XML Schema definition file. The Task Pane provides step-by-step guidance to Office users and walks them through the creation of documents that conform to the schema.

Schema validation, along with promised document transformation support, indicate that Microsoft is serious about XML in Office. Office will not only warn users about wandering outside the lines as they enter data, it will ensure that saved files are encoded in a predictable format. External applications will be able to create, interpret, and modify Office 11 XML documents. As encouraging as this is, there is a drawback: Schema definitions aren’t easy to create. But if you feed Office an XML document, Office will attempt to divine a schema from it. When combined with the Web services hooks Microsoft has already worked into the product, Office 11’s XML capabilities will help eliminate old barriers to integration with external applications.

Mac OS X

It’s inadequate to cast OS X as an upgrade to Mac OS 9. Apple loyalists can point to a host of OS X features that actually debuted in the Classic versions of the Macintosh operating system. But that’s as relevant as tracing Windows 2000 features back to Windows for Workgroups. Where most business customers are concerned, the Mac was purely a niche and consumer player until OS X was introduced.

OS X scored instant credibility by combining best-of-breed technologies. The mature and rock-stable Mach kernel, the BSD 4.4 libraries and commands, the OpenGL graphics engine and the NeXTStep API made OS X a known entity from the day of its release. The milestone release of Version 10.2, and the significant updates (10.2.1 and 10.2.2) that followed, proved Apple’s commitment to building out OS X’s features and responding to customer concerns. But perhaps most impressive is Apple’s concurrent publication of most of OS X in open-source form as Darwin. Darwin is quickly joining Linux and BSD as must-hit targets for open-source development and is rapidly rising in popularity among commercial developers.

The disruptive impact of OS X as a client environment is overwhelming and undeniable for any who give it a fair shake. OS X put Apple on the map as a player in the general IT market, enlivened the slumbering notebook market, rescued the PowerPC processor from obscurity (if not extinction) and gave developers of all stripes a compelling showcase for their talents. OS X Server’s disruptive power is not yet as apparent, but Apple has staked out flat-rate pricing, low hardware operating costs, and simplified management as differentiators. It will take another year or two for Apple to turn its server products into the kind of hands-down category winners that its notebooks and client OS are. Maybe that gives Microsoft and open-source OSes time to get ahead of Apple. Other contenders will have to run a lot harder to make the finish line now that Apple’s in the race; so no matter how it plays out, IT organizations and users benefit.

Open source

OSS (open-source software) has traditionally been a playground for geeks. Stereotypical free software is technically rich, poorly documented, completely unsupported, and despised (or laughed at) by commercial vendors. OSS has always disrupted the market by supplying zero-cost alternatives to expensive applications and shining a harsh light on vendor licenses. But the knowledge required to participate in open source — as a basic consumer, not as a developer or an expert — created a barrier to its use in IT. OSS developers have shown little interest in making their work accessible to less-knowledgeable users.

That was before the recession. IT’s interest in OSS has spiked of late, driven by companies’ need to trim costs and to push back against rising fees from vendors such as Microsoft and IBM. As IT jobs became scarce, OSS developers turned their projects into portfolios for their skills, raising the quality and completeness of their work dramatically. Vendors started looking to the OSS community as a distributed think tank. It’s the best place to shop for sharp talent and to road-test new concepts.

IBM, Sun, and Apple discovered that they could get a huge head start on new projects by incorporating open source. Commercial vendors trade hardware, money, jobs, and visibility for OSS’s proven code. IBM’s storage appliances, Sun’s desktops, and Apple’s flagship operating systems grew out of OSS/vendor partnerships. These pairings reduce vendors’ development costs and shorten their release timetables. Extending open source’s collaborative model to commercial players benefits all involved. When you mix the ingenuity of thousands of OSS developers with vendors’ skills at packaging and support, the result is software that IT can trust, understand, and afford. That’s the best kind of disruption.

IT executives can cheer about open source, but so can administrators and individual users. Linux, FreeBSD, and OS X will simplify access to the Internet’s immense library of free software. For example, the Fink package manager available for download from Apple’s site — Fink is free and was developed independently — provides Mac users with a browsable catalog of nearly 2,000 free software titles. With a single command, Fink will locate, download, compile, and install any program the user selects.

Open-source developers are working to expand the list of precompiled (binary) applications for OS X and other open operating systems. If a binary exists for a program, the package manager will skip the lengthy compilation step. But even if an application must be compiled before installation, the package manager takes care of it. In many cases, no programming skill or knowledge is required. All it costs the user is time.

As OSS takes on more of the qualities of commercial applications, it will be easier for IT to fit open source into strategies that include vendor-supplied solutions. Knocking down the knowledge barriers with projects such as Fink will expose open source to millions of users that don’t know it exists.