Mission Control

Beyond NeXT's sloganeering, customers are deriving real advantages from mission-critical custom apps

by Eliot Bergson

When Karl Fankhauser looked around the sprawling Pinole Point Steel plant in Richmond, California, shortly after getting his first shipment of NeXT machines in October 1991, he saw manufacturing and administrative needs that couldn't be met by conventional software. So he developed his own.

As manager of information systems, Fankhauser put together a staff of three programmers and built a process-line scheduling program that helps forecast and manage production rates, galvanizing, organic coating, inventory, and a host of tasks particular to Pinole Point's specific line of steel products. The programming team is now developing other applications to tie together all of Pinole Point's 55 NeXT machines, in every major department, to forge a completely integrated computing environment for its business.

The business of custom apps is business. Fankhauser and his team aren't just developing productivity tools for individual workers; they're building a suite of custom apps to house an entire steel plant's operations. They're using software to comprehensively support Pinole Point's business mission. They're developing the mission-critical custom app.

Pinole Point is not alone. Many of NeXT's biggest customers are using NeXTSTEP development tools to build in-house software to automate their business processes. But as more and more firms develop mission-critical custom apps, their strategies, problems, successes, and results vary. Important questions arise about what terms like "custom" and "mission-critical" really mean, how development should take place, the role of user feedback, and the mandate of information systems (IS) departments. The only common denominator for developers using NeXTSTEP seems to be enthusiasm: "No comparable environment is as user-friendly. We all think it's the greatest thing since sliced bread," says Fankhauser.

An app is an app is an app?
Endorsements like Fankhauser's are probably music to the ears of Steve Jobs, who heralded the era of "mission-critical custom applications" in his keynote at NeXTWORLD Expo in January. By spring, NeXT was pushing NeXTSTEP's friendly-yet-powerful development environment as its major marketing tool, taking out full-page advertisements in the Wall Street Journal urging companies to "Build a Weapon Your Competitors Can Never Copy." And, unlike machines from NeXT's competitors (which are being positioned to push the message of custom-applications development), the black box has that capability now.

Mission-critical apps can be so strategic that many firms, especially in the highly competitive financial industry, won't even talk about them. One IS director at a bank says he's "afraid of losing good programmers to other shops." Another was more blunt: "Most of this work is proprietary. We don't want our competitors to know what we're doing."

Secrecy about custom apps only underscores the fact that many corporate developers, in a variety of industries, are heeding Jobs's call. Traders at oil-commodities giant Phibro Energy are starting to use a real-time risk-management system developed in-house. Phibro has also spun off a software-development company, mc2 technologies (see "NeXT buyers become sellers," NeXTWORLD Extra, March 1992).

Peter Armstrong, vice-president of mc2 technologies, says that his programmers had to develop custom software because of the unique nature of the oil-trading business. "There was nothing out there to cover the energy market. The physical aspects of the market � shipping, pipelines that break down, even political developments � are so complex that there were no apps to pick from, on any platform," he says. In six weeks, he and five other programmers developed four applications: an option-quoting tool that gives clients instantaneous risk assessment; a graphing program to compare re-source energy products in thousands of markets over different seasons and price series; a multimedia phone book and calendar; and a trade-capture app that has since become the cornerstone of the large risk/management system being marketed by mc2 technologies.

Armstrong and his team built "hundreds" of objects during the development process that can be reused and combined to produce more applications. The contact manager, for example, was an eye-opener for Armstrong because he "looked at the shrinkwrapped products already out there, but again, none of them met our needs. They weren't energy-specific, and even though we could have saved time and bought one, there was only one or two on the market. So we leveraged our object work and built our own," he says. His team was able to not only customize the ideas behind contact-management software but to produce a robust app that can incorporate spreadsheets, images, word-processed documents, and faxes � all with access to the risk-assessment tools of other in-house software and features, like the Colors panel, that are inherent in the NeXTSTEP environment.

Reusing objects offers companies the double benefit of tying together various parts of their businesses with common, underlying functionality or even marketing their finished software, while in-house programmers can easily customize versions of the same software � changing interfaces or operability, say, for back-office versus front-office users.

This offers a level of functionality that shrinkwrapped apps can't approach, according to Eric Bergerson, managing director at Objective Technologies (OTI) in New York, the first commercial vendor of object-extension palettes for Interface Builder. He likens objects on the NeXT to a Lego set that contains basically two kinds of pieces: blocks, and doors and wheels. The blocks are simple objects, while the doors and wheels are higher-order objects. "You can use pieces of both sets to build a helicopter, for example, and then dismantle the helicopter and use those and other pieces to build a four-wheel drive vehicle," he says.

Another major feature of in-house development is the ability to prototype an application and get instant feedback from the users who will be using the software. For the front end to the Lexis database of court decisions developed at Marger, Johson, McCollom & Stolowitz, a patent-law firm in Portland, Oregon, programmers regularly quizzed workers about the work in progress. "We distributed the .nibs and said, 'Fire it up and try it.' There wasn't any database behind it, but they could run the interface and get a feel for how it works. They could do text entry, play with tab fields. It was real software architecture at the user level," says Gregory Miller, president of Inherent Technologies, a software company spun off from Marger, Johnson.

The team at mc2 technologies took a still more interactive approach. Developers sat down with users at the computer and constructed the user interface right in front of them.

"Developers were happy because they wouldn't have to tweak it later on. And users were very interested in having apps built for them and helping out. I've never seen our users so interested in computer systems before," says Armstrong.

Business as unusual
Changes in how software is developed may be calling into question the very roles of workers and information-systems specialists, according to Assistant Professor Rajan Srikanth, of the Walter A. Haas School of Business at U.C. Berkeley. As firms began to purchase mainframes and minicomputers in the 1970s, an "ivory tower" IS department would write software for the company. Workers had to learn the new programs, which were filled with arcane and arbitrary function-key commands, and had little chance to offer feedback, even after the app hit their screens. With the advent of the personal computer, though, "end user" computing became the norm, with shrinkwrapped applications offering workers the chance to increase productivity � and better understand computing through enhanced graphical interfaces. "What's happening now is a re-emergence of centralized IS, but in a different form. The department may not become a centralized power concern, but the thinking about its function will be," says Srikanth.

Office workers are driving the current trend toward custom programming, he explains. As computers become more than just tools to enhance productivity, they create opportunities that never existed before. Srikanth cites the increased use of groupware, and how it allows users to work with others in real time, as an example of how a business can gain unexpected competitive advantages. "The democratization of computing that occurred during the desktop revolution is continuing to evolve," he says.

OTI's Bergerson, who has designed apps and interfaces for several companies in addition to his pure programming experience, says "you have to remember that workers in an office have never had the chance to tweak either a big program from IS or a shrinkwrapped app. They had tools to help them do little tweaks, but the programmer still produced professional apps. Now workers can get their hands on higher-level doors and wheels because of the NeXTSTEP development environment and user interface. They can learn Interface Builder if they want to, build prototypes on their own, then go to IS and get it done in code. We've distributed the power of the IS department."

And as workers become more proactively involved in the development process, explains Inherent's Miller, IS departments will become "object- constructers." Users will outline their needs for an object with specific methods and data, and the IS department will build it. Users will then build from the new object, while programmers will in turn see new uses for these objects. "Databooks for objects will come into existence," says Miller.

Critical path
For Pinole Point Steel, this new approach to programming has allowed the IS staff to build applications that modernize a business based on both old and new technology. "We're seeing a basic change in IS, and with the level of technology we've purchased, we can have a very small staff. We're pushing our IS people to give users the ability to navigate around the interfaces. I'd like to see natural-language queries [to our Oracle database]. If they don't exist, we'll create them," says Fankhauser.

"For us, the mission is to increase information. We supply a high-quality product, and we can't improve on that. What we can do is improve real-time information about our product. When customers call for information about an order, we can get it to them immediately � and we give them a lot of information," he says.

For many developers, the ability to reuse objects and prototype interfaces, when combined with the changing roles of IS departments and end users, has allowed them to program for specific business operational concerns rather than merely increase worker productivity. This is the difference between any custom app and one that's "mission-critical." Businesses must look at "core competencies and incompetencies and decide how to best leverage those core competencies and best improve those core incompetencies," explains U.C. Berkeley's Srikanth.

"Mission-critical means 'I've got a job to get done.' Except for this app, the mission is not going to get done. It's also called 'critical path' in engineering," says Inherent's Miller. With that in mind, Inherent is now developing the first docketing tool and client-database model for the legal profession.

In a similar vein, developers at Swiss Bank Corporation/O'Connor Services (SBC/OC) are building a complete financial-trading system to swap debt and sell bonds around the world, in seconds. At SBC/OC's offices in Chicago and London, traders use banks of phones to complete up to ten deals a day. A government, for example, may want to swap debt with a foreign bank or the International Monetary Fund for some current fixed rates. Deals routinely involve hundreds of millions of dollars over a 20-to-30 year period, according to Vice-President Bill Martin. "We really focus on representing our real phone world. Our business is driven by the software we write," he says.

Moves toward custom programming in critical business areas signal a shift in how computers are viewed in business, as well as an improvement in how businesses view their missions, according to Srikanth. Since the computer is no longer seen as just a productivity tool, people no longer view technology as a singular solution. "

'Mission-critical' is technical jargon, and certainly different for different people," says Srikanth. "Businesses must see in what ways it changes how they do business, not just how they compute."

A future or the future?
It's evident that programmers at NeXT sites are trying different strategies to secure a foothold in the world of custom-application development. With so many ideas emerging, some critics are beginning to issue warnings. Until the NeXT market stabilizes or grows, developers and businesses may be only exploiting a "closed-niche opportunity," according to Rikki Kirzner, senior industry analyst with Dataquest in San Jose, California.

"A lot of folks feel they can get return on investment, but they're not looking at the long term. Lots of things can change," she says.

Change does seem to be the only constant, with the diversity of development strategies a signal of either health or disarray: Businesses are leveraging their development work by launching spin-off software companies; Educational researchers, like UC Riverside Chemistry Professor Kurt Monnig and graduate student Jeff Hagen, who are developing drivers for capillary electrophoresis machines they're building from scratch, are planning to market their software; Williams Telecommunications, a giant fiber-optic research firm, is building a management system for a public Asynchronous Transfer Mode network, the next generation of data communications; Trirex Systems, NeXT's new East Coast training partner, has developed a suite of customizable "installation apps."

Despite the secrecy involved, rumblings from Wall Street can be heard as firms try to match the competitive advantage grabbed by custom-app developers at Phibro Energy, SBC/OC, and First National Bank of Chicago. And government agencies certainly aren't using their black boxes as paperweights.

On the support side of the equation, a host of new consulting companies are marketing their programming expertise, so that every firm that purchases NeXTs won't be forced to have in-house programmers. Even established third-party developers are beginning to see the power behind mission-critical-applications: RightBrain Software, for example, has begun to bundle source code with several of its utility programs, in order to allow for customization by the end user. And Stone Design's founder, Andy Stone, says he'll "put in real clear APIs [on new software], ways that end users can roll parts of our app into their app."

The mixing and matching of objects, code, and business missions will only continue. In the meantime, users may start seeing their black boxes as Inherent's Miller sees his: "My NeXT is my information appliance. It's my data cuisinart."

Eliot Bergson is the associate editor of NeXTWORLD.