The very first database application I ever developed was a small departmental system that tracked invoices for temporary staff. I programmed this "database application" in Lotus 1-2-3 version 1.0a using macros. When I was training my successor for that position, he asked why we did not use a "real" database manager, such as dBASE II. I explained: "That's what our department has, so that's what we use." I would not be surprised if this rationale drives the choice of tools for many projects today, including much larger and more complicated database management applications. Using less than optimal tools on hand for any kind of project is an age-old problem that is not likely to go away.
The multifunction integrated package strategy popularized by Lotus 1-2-3 reached its pinnacle with Lotus Symphony and Ashton-Tate's Framework. But cramming multiple functions into a single package proved awkward and unwieldy; it has been displaced by suites containing separate specialized programs. Yet dozens of client/server application development tool vendors are retrofitting their existing tools to support new Web technologies such as Java and HTML. The allure of "one size fits all" integration can be a powerful magnet for vendors.
After dominating end-user applications with Microsoft Office and then extending the suite strategy with Back Office for system administrators, Microsoft Corp. has now bundled many of its application development tools into its new Visual Studio suite. The entry-level Professional Edition of Visual Studio is a weighty package that includes Visual C++, Visual Basic, Visual J++, Visual InterDev, and Visual FoxPro. The kitchen-sink Enterprise Edition adds fully functional copies of SQL Server and Transaction Server (the license covers development use only), plus Visual SourceSafe, Microsoft's version-control system. The new Microsoft Repository is in there, too. You even get the coveted Developer Network (MSDN) Library CD-ROM. Stir in a third-party testing program (Microsoft recently sold its Visual Test product to Rational Software Corp.) and a modeling tool, and you have a pretty comprehensive tool chest. You might also want to order another hard drive.
The technical rationale for a development tool suite seems compelling. Building a modern application increasingly requires the ability to work with more and more technologies such as Java, HTML, scripting languages, C or C++, and a 4GL such as Visual Basic. Why not get one of each in a package deal -- and if it saves a few bucks, that's even better. A shared Integrated Development Environment (IDE) and the promise of cross-tool integration is appealing. Getting all of these tools from one vendor might diminish finger-pointing when cross-product technical support is required. Sounds good so far.
So where is this road ahead taking us? You don't need a Harvard MBA to figure out that suites are a strategy to dominate multiple product lines at one time. Before Microsoft Office obliterated its competitors, Microsoft and other vendors waged separate battles for market share in word-processor, spreadsheet, presentation graphics, and other product lines. It was not uncommon to find a corporate standard that mandated diverse best-of-breed products from separate vendors. Today it's much rarer to find a company using Microsoft Excel and Corel Corp.'s Word Perfect, or Microsoft Word and Lotus 1-2-3.
Today's application development tool market is where the office tools market was years ago. Microsoft's Visual C++, Visual Basic, and Visual J++ compete as separate siblings fighting it out with Borland International Inc.'s C++, Sybase Inc./Powersoft's PowerBuilder, Symantec Corp.'s Visual Café Pro, and a variety of other competitors, including hundreds of tiny Web development tool startups. If Visual Studio succeeds even half as well as Microsoft Office has, it will be at the expense of a more competitive and innovative tools industry. When so many tools come in one box, many users won't make the effort to obtain and try out competing tools from other vendors. How many vendors still sell single products that compete with Microsoft Office applications?
Suites such as Visual Studio also highlight the continuing tension between learning new skills and maintaining high quality and productivity. Bundling five or more tools doesn't reduce training burdens. A common IDE helps, but mastering language nuances will always take time. It's fun to learn new tools and technologies. But real productivity and the quality of your code usually decline while you scale the learning curve.
I'm no longer a full-time developer, but I still spin out a little application or two now and then. Of course, I'm now way beyond Lotus 1-2-3. Today I build my database applications in Microsoft Excel. Microsoft Word tables are too limiting.