Tale of Two Markets

Let's start by savoring the irony that the biggest thing ever to hit business — the Net — was the creation of hackers working with government money. They didn't do it for the money, and they didn't do it to make money. They did it because it needed to get done, and because doing it brought a priceless kind of esteem: the respect of their peers. Crafting the Net was a good and noble Work that still goes largely uncredited outside the circle of peers who best understand what really happened. Given the importance of that work to the economy, it's a wonder these guys don't have statues on Wall Street.

Now the stock market is going gaga over Linux for the same reason it went gaga over the Net: it's good for business. It doesn't matter why to the market. Putting Linux to work is kind of like putting an e in front of your idea or a .com behind your company name. It's a grace. A particularly lucrative grace, it now appears.

But just as the Big Boys didn't understand what that .com really meant at first, they don't understand what Linux and open source mean, either. They thought the Net was just a 10 million channel TV with a keyboard instead of a remote control. Likewise they think Linux is just a 10 million hacker version of Microsoft with bloated attitude instead of bloated software.

Not that the "free software" community is especially understanding about Business. Scott Lanning has this to say about Linux Journal for a Seattle Times story: "It's proprietary, so it means nothing."

Makes me wonder: how are we "proprietary?" In Open Sources: Voices from the Open Source Revolution, Richard Stallman says "The 'Linux' magazines ... are filled with advertisements for proprietary software that works with GNU/Linux. When the next Motif or Qt appears, will these magazines warn programmers to stay away from it, or will they run ads for it?"

Well, probably both. Why shouldn't we give Troll Tech a way to say good things about Qt tools? What's so wrong about that?

Answering this question will, of course, bring up blood-boiling moral questions that find their most concrete expressions in licenses and copyrights (also copylefts) that are strange and baffling to those who never heard of "open source" or "free" software until confronted with the need to use it. What's so "free" about something that comes with so many restrictions — and with so many people who are quick to flame you for breaking rules only they understand?

But Business has a real demand for what Eric Raymond calls "software that doesn't suck," and for the people who make it. Likewise, those people have a real demand for good development tools. And not all of those tools are going to be free or open source. Qt is an old example. The new example comes from Inprise (formerly Borland, also known as Borland/Inprise).

This past July, Inprise ran a survey of  developers visiting various Linux sites, including Linux Journal. Of the 24,000 unique responses, about 6,000 called Linux their primary development platform. Naturally, the leading development tools for this group were GCC/EGCS (47%) and Emacs with GCC/EGCS (27%). Yet a majority of those same 6,000 developers were willing to pay for commercial Linux development tools: 27.9% said they would be willing to pay $300 or less, and another 37.2% said they would be willing to pay $100 or less. Only 21% said, "Nothing, it must be free."

And what kind of tools did they want to buy? By class the favorites were "Rapid Application Development IDE (RAD, visual development)," at 53.6% and "traditional development IDE (Integrated editor, compiler, debugger)," at 35.7%. As for particular tools, first choices were "C++Builder (C/C++ with RAD)," with 29.6%; "New IDE that works with existing standard Linux tools," with 24.2%; Delphi, with 19.2% and "Borland C++ (C/C++ without RAD)," with 13.3%.

So Inprise has responded to demand, with Kylix, the code name for "a high performance native Linux development environment " a Rapid Application Development environment that will support C++ and Delphi development," says Michael Swindell, who heads the project for the company. It's a major effort. "We're developing a visual component framework to radically speed and simplify native Linux development. Graphical, database, GUI, Internet, multi-tier development will all be completely visual and component-based, using the Borland VCL and two-way tool technologies."

Just like Delphi and C++Builder for Windows. "It's architecture is a component abstaction directly over the native environment. It is derived from the VCL architecture the current Delphi and C++Builder products are based on, so the component APIs and methodologies will be very similar. So familiar that if you already develop with Delphi or C++Builder for Windows, it will be very easy to do Linux development from day one. Essentially, you won't have to learn the complex details of the underlying architectures, though they'll still be completely accessible, making new Linux development much faster and porting far easier." Swindell adds.

Now here's the really tricky part, because this thing is cross-ethos as well as cross-platform. "We are not developing Kylix as an open source project, but we are investigating which parts of Kylix lend themselves to being open source. Right now Kylix will allow developers to create open source applications. But we have not determined which, if any, components in Kylix will be open source. Our goal is to enable Kylix to develop both proprietary and open source applications, because there will be markets for both."

As I write this (in late September), Corel is getting beat up for putting proprietary boilerplate on the beta version of its new Linux distribution. No doubt Inprise will run through a similar gauntlet — all the way to the market. In fact, Swindell expects it. "This is a discovery stage that a new market has to go through. You've got commercial vendors coming into the open source space, and open source vendors moving into the commercial space. It's a merging of very different business practices. We have to go through a lot of head scratching before this settles out."

How will each model subvert the other? Consider the parting words from Michael Swindell: "We need the extremes. These are the people fighting for the noble causes. We can't discover open source standards and policies without them. And frankly, open sourcing is not a closed question for us. We'd welcome the open source community's involvement in our own development. But we've got fourteen years of intellectual property and patent accumulation, plus legal and shareholder interests to look out for. There are extremes on that side too. The best we can do right now is answer the market and talk to everybody with an interest in making this work."

The default assumption is that the most powerful interest — shareholders — will want Inprise's source code to stay closed. But it's not hard to imagine investor pressure on traditional software companies to free their source code, just as the same investors pressured ".com" companies to free their content.

Freedom is an efficiency that drives value. Isn't it fun to watch this new software business teach Adam Smith's lesson, one more time?