Issue 1-17, April 3, 1996

Be Engineering Insights: Of Base Classes and Flexibility

By Erich Ringewald

After you've spent a few years creating a new platform behind closed doors, huddled like mad scientists, it's possible to lose perspective. That disappears quickly when you ship a few hundred machines out to developers who simply want to develop on a great machine. They start to pick it apart like crazed weasels, and I mean that in only the most positive sense. The last few months of supporting the pioneer Be developers has been very interesting.

The C++ language is one subject that comes up quite a bit. Some developers love the fact that we've decided to use this language standard. Others appreciate that we have avoided the language's more "esoteric" features (multiple inheritance is frequently mentioned here). Still others (actually only one or two) have said that although they love the BeBox and its kits, they just can't program it because C++ nauseates them. A visceral reaction to semantics!

Admittedly, there is a certain amount of religion in these positions, but one thing is true: C++ is not known for its dynamism. Between static binding, the base class feature, virtual tables, and no garbage collection, C++ is a bit of a stubborn old aunt who refuses to leave the house. For us, this presents some problems and challenges, but ones which I think can be managed if we all modify our mindset a bit.

With the older platforms whose limitations we are trying to overcome, software (both application and system) has been a very static thing. It takes about four years for Microsoft and Apple to develop and publish a new release. Application software is then stuck on floppies or CDs which spend an average of nine months in the distribution channel. If we are going to be able to continue to outrun those platforms, we need to have a much more flexible attitude about software updates and distribution.

Such flexibility in thinking will help us together overcome some of the static limitations of C++ without locking ourselves into an API, virtual table layout etc., for all time. But it's not just to make dealing with C++ easier that I recommend this, it will be better for our customers as well. They ask you for new features in your application, you in turn ask for new system software features to facilitate them, we implement them, perhaps shifting around a class layout. We ship you the new version of BeOS, you implement the new features, we ship the new software together, and the customer buys the upgraded application from you and thinks you're the most responsive application developer they've ever seen.

So how often can we do this? Every four months may be a bit too frequent—but it will certainly be much more often than once every four years. The improved communication and distribution mechanisms afforded by the Internet will help us show ourselves to be much more responsive than the older platforms.

Be Developer Profile: AMP Productions, Inc

The BeBox will shine in multimedia,” says Tom Ierna, president of AMP Productions, Inc. Based in Florida, his 5-person company specializes in 35mm slide presentations, video, computer-based kiosks, and live entertainment, serving the education and business markets. When four thousand members of local student clubs convened in Florida, AMP was there shooting slides and video, which they edited on-site to produce and deliver a show on the last day—much like a slide show on the last day of summer camp.

Customers today want integrated video, computer graphics, MIDI, and digital sound effects in their slide shows,” says Ierna. “The cost of the hardware, software, and expertise to produce that kind of presentation is phenomenal: a 15-20 minute show can cost up to $30,000. You almost want to ask customers to hand over their checkbooks!

As a small company, AMP is always on the lookout for ways to make their multimedia productions easier, faster, and less expensive. And that's exactly why the BeBox caught their eye. “Out-of-the-box hardware extras are what made us choose the BeBox,” Ierna says. Especially the MIDI and GeekPort capabilities, but also the PCI bus architecture, the CD-quality stereo sound, and the slew of expansion and connection options. “With the BeBox, we feel like we're working on the cutting edge instead of on the 10- to 20-year old technology that most of today's high-volume computers are based on.

AMP has both short-term and long-term plans for the BeBox. Short-term, they'll use their Be application internally to produce the slide presentations, with the BeOS™ GUI replacing their command line-driven programming language (which has been adopted as the industry standard). “We'll also use our application to control the slide shows: the slide projectors will be controlled by the BeBox (most likely through the GeekPort), our audio will be generated on the BeBox, and we'll use a control signal from the BeBox to sync the video.” Long term, they hope to see digital video editing and output on the box, providing a truly integrated solution.

Ierna doesn't think it will take much to attract multimedia developers to the Be platform. “Mention true multiprocessing, true multitasking, multithreading, the expansion and connection options that the box provides, and they'll froth at the mouth.

AMP expects to release their Be application in early 1997. For more information, send an e-mail message to


By Jean-Louis Gassée

This week we lost a big man, David Packard, one of the Founding Fathers of Silicon Valley. The mourning has less an air of sadness than one of celebration for we are remembering an accomplished life. There isn't much I can add to the well-deserved tributes to his business and philanthropic achievements. "The HP Way," a charmingly unassuming book, details the philosophy behind the business "Bill and Dave" built,perhaps the most enduringly successful company in our industry, good to its shareholders, customers and employees. I know, I worked there.

Twenty-eight years ago this coming June, Hewlett-Packard offered me what proved to be the opportunity of a lifetime. At age 24, I was coming out of what is tolerantly referred to in California as my psycho-social moratorium. Meaning my business experience consisted of a string of disconnected jobs—descriptions of some would be out of place in this family publication—in the food, beverage and entertainment, pharmaceuticals and insurance fields. I had no idea what a resume was. So, when I saw the recruitment ad specifying a math or physics degree, selling experience and fluent English, I wrote a two-page letter explaining why they should hire me. They did, in spite of my broken English. My job was to launch their first desktop computer in the French market.

Looking back, I shudder sometimes. Placed today in their situation, would I hire myself? Joining Hewlett-Packard was a revelation. I came from a culture of distrust and hostility between management and workers. At HP I grew in an environment where I was trusted, given a degree of freedom and responsibility I had not encountered elsewhere, and found an incredible patience with my combination of inexperience and strongly held views. I flourished and rose in the hierarchy as HP became king of the hill in the desktop computer business, before it missed the early phase of the PC revolution.

To this day I'm indebted to HP for its kind of schooling. It was so convincing I came to think every US company conducted their business in the same humane and efficient way. I was brought back to reality when I joined another school—of hard knocks this time—at a very East Coast minicomputer company. A few years later I got an opportunity to practice both kinds of schooling when I started Apple France. Our office was across the street from HP so I could hire a few good men from the old company without disturbing their commute. We did well and, in obeisance to the Peter Principle, I was transferred to this country and promoted to a job for which I had no experience or formal training.

After another five years of schooling in the ways of Corporate America, I started Be. The HP connection helped in small and big ways: When HP's credit department was unwilling to ship a logic analyzer to Steve Sakoman's garage—a small but frustrating problem—John Young's (HP's CEO at the time) secretary quickly cleared the path. The Big Problem was fund raising. When we started Be, our idea looked even crazier than it does today. Windows and Mac reigned supreme, investors in NeXT and Momenta had been burned and Go didn't look great either. Kleber Beauvillain, CEO of HP France and an early Be investor with other HP Europe luminaries, made the introduction to Credit Lyonnais. Their venture arm, Innolion, became the lead investor in our first round. The company would have probably never gotten off the ground without that connection at a time when most US venture funds (with one exception, Newtek) found the idea too risky—or "too interesting" as the dreadful euphemism became familiar to us.

As we make the transition from a group developing a product to a real company, one with developers, business partners, and customers, I too mourn the passing of a great man. I do it with admiration for a life well lived, with gratitude for the opportunities and the schooling, and with the hope I'll honor my debt in building this business.

Creative Commons License
Legal Notice
This work is licensed under a Creative Commons Attribution-Non commercial-No Derivative Works 3.0 License.