Open Source

Ever since my daughter Robin was born, I've been covering education technology.

It has been a history of failure.

Back when Windows was rendering everything obsolete every three years, technology coordinators in education were dead men (or women) walking. As soon as they got systems ready for training teachers, the systems were obsolete.

This is no longer the case. Computers are cheap. Since their primary use is to access the Internet, they remain useful for years. But only if these assets are managed. And solutions like Tivoli, are not cost-effective in the K-12 market.

So schools' technology coordinators are still dead men (and women) walking. Especially when they tell the bosses that the hardware and software they're buying is just 15% of the money they need to spend.

Solutions are finally coming from the world of open source. From Open Country, CEO Michael Grove told me. "We do system discovery and systems management. We can do remote monitoring, asset discovery, diagnostics, repository management, provisioning, deployment, security updates and scheduling."

With Open Country in the hands of a vertical market vendor, school networks can be managed remotely for the first time, for a reasonable fee. School computer budgeting becomes predictable.

Grove has been working with Intel to deliver these kinds of cost-effective solutions to India under Linux. "We can sign channel partners to roll trucks. Now you can just press a button on our graphical user interface and reprovision without having to know anything. In the background our management system is doing the lifting."

Great. When will it be available here? "We’ve found a group who is working with lots of school districts and is putting together solutions. We’ll have Windows by year-end."

Robin graduated high school last month.

Microsoft has left the discussion.

The vaunted open source blog of Jason Matusow has not been updated since May 24. And Robert Scoble, the firm's most recognized blogger, has left for a start-up. He leaves at the end of the month, but his blog has already moved.  

Is it wise for Microsoft to leave the discussion, and return to its top-down ways, just as open source is heating up?

Even with the presence of Matusow and Scoble Microsoft had little support in the open source world, so in some ways it doesn't matter. At some point you are what you are, and Microsoft will never be an open source company. It's too old, has too much history, and has been too successful in the closed-source world for that.

Besides, you figure, the hottest company in tech right now is Apple, a closed-source company. Where is the upside? Doesn't engaging in open source discussion just reflect weakness? You ever seen an Apple exec blogging?

My view is that Microsoft was never going to give more than lip service to openness, so why bother? If all you're going to pay me is lip service, frankly, I'm happy to decline the honor.

To get, first give.

It's among the oldest Internet values. But having gone through the dot boom, the dot bust, and beyond, VA Linux founder Larry Augustin says it's still the key to success.

I caught up with Augustin, who left his full-time role at VA in 2002 (he's still on the board) and now makes his living as a venture consultant and board member in the San Francisco Bay Area, between meetings and found his ideals are intact. Among the open source start-ups he works with are Hyperic, Xensource, SugarCRM, Fonality, and Medsphere.

In Augustin's view open source development became a necessity in the 1990s when the cost of marketing a program came to exceed the cost of creating it. "My favorite is Salesforce.com. In 1995 they spent under $10 million in R&D and over $100 million in sales and marketing. That doesn’t’ work.

"Open source enables people to reach all those customers. It’s a distribution model. The people who create great software can now reach the rest of the world."

Businesses get the most protection from the GPL, he insisted. "They get protection from competition." The license's insistance on reciprocity means no one can take the code you wrote, tweak it, then compete with you.

As to VA Linux, which went south after trying to make it in hardware, "There was a window where the world wanted a hardware supplier who could deliver high density rackmounts optimized and supporting Linux. When we went through the dot com crash the customers who wanted that went away." The world has changed since then, he added. "It’s incredibly tough right now to be a hardware supplier."

Perhaps his most interesting statements concerned the "threat" of competition from India and China. If they want traction in the software market they, too, need to give a little bit.

"I still don’t see many people from India and China contributing to open source. It’s when the people in those countries get into the process that you’ll see things. It’s not going to happen until you see the contribution level increase."

One point that bears repeating following our great discussion about proprietary business models the other day is the idea of sunsets.

The Founding Fathers knew that even the best ideas had a limited shelf life. The patent system gives protection only for a limited time, and only for the original idea. (Others can use your design to do something better.)

The copyright system was also designed to provide limited protection for a limited time.

The idea is that the incentives now called "intellectual property" are designed to be limited, in time and in scope, so as to encourage the production of more. It's an idea today's holders of copyright and patent rights conveniently ignore.

The usefulness of software ideas is also limited. Competition often brings better ideas. Over time, prices should go down for the basic stuff.

In the PC era, however, they did not. This was because Microsoft controlled the bottom of the stack. Microsoft could continue charging monopoly rents on basic ideas, could even devour the application niches above its software and claim monopoly rents there, because it controlled the bottom of the stack.

Open source is not like that. What happens is that, over time, the need for support on a new package declines, and the revenues to be expected decline as well. As the operating system and basic applications become standardized, less support is needed.

Sunset is finally achieved.

To compete against this model, proprietary vendors must constantly create new capabilities, valuable capabilities folks are willing to pay for, and that they will want support for. Otherwise they should expect prices to fall, and their revenues to fall. This is how open source vendors work.

I'm sorry if Microsoft, and its supporters, do not like that idea. But that's just business evolution in action. And after some decades in stasis, it's finally operating again.

Thanks to open source. Enjoy your sunset.

Former senior engineers at Covalent have announced an open source business for Hyperic HQ 2.7, their IT management product.

The older 2.6 version is supported by Covalent and it supports PostgreSQL .

CEO Javier Soltero said his team built the product at Covalent, then bought it from them as it concentrated on Apache.

"They had sold it as a Web management solution. But we’d designed it as an adaptable, cross-functional management tool that could manage any platform," he said.

"So we are bringing this to market with a different focus. We’re changing the licensing model from old-school enterprise licensing to per-machine subscription, which includes every aspect of the product."

The open source model is being announced just as Hyperic is taking its first venture capital investment, a good indication that VCs are looking for open source opportunities.

"We were pretty fuzzy in the funding announcement talking about what we would do with the money," Soltero admitted. "The real announcement is we’re going to open source the platform, under the GPL."

Soltero says he sees JBOSS as a model, and has been offering Hyperic as the management platform with JBOSS since last summer. "Our JBOSS work has been a very good partnership and we hope to replicate it. We want to do major league level business with open source."

Open source cynics, please take note.

Everyone knows AJAX (Asynchronous Java and XML) is the Next Big Thing.

By combining programming with tagging, you can do amazing things. You can embed applications inside tiny bits of hypertext, and get a lot done in the background with minimal user action.

Right now there's a rush to make all this open source. We know Java itself is moving toward an open source model. IBM is now pushing heavily on this front, proposing an open source AJAX project to Eclipse, and donating some code.

This is where it gets hairy.

Everyone wants their code in the standard platform because that makes them the author of the standard. But for a platform to be a standard, it has to be standardized.

There used to be two ways to create a standard. There was the Microsoft way and the committee way. Windows is a standard done the Microsoft way. WiFi  is a standard done the committee way.

Trouble is, the commitee way does not move quickly. We have seen this with WiFi, where "preliminary" designs for new standards are out before the ink on those standards is dry.

This is what we need to do with AJAX. The question I have to ask is, simply, does the committee have the resources to address this quickly and keep AJAX uniform?

I hope so.

Reports that Microsoft and Adobe are breaking out the lawyers over the PDF format is being spun in many ways. As an antitrust battle. As a feature battle.

Is Microsoft trying to run down Adobe? Is Adobe being too close with PDF?

But one piece of spin seems incontrovertible. It's another loss for the proprietary business model.

This sort of thing just does not happen in open source. When file formats are open source, any other project who offers support for it is offering just that — support.

When money and market share aren't at issue, cooperation becomes assumed. If everyone in the open source world is using the same file format that's not a monopoly — it's a standard.

The proprietary world cannot work in this way. It is always about my property vs. your property. Only the rhetoric gives the user service, and then it's lip service.

This is why the replacement of the proprietary software model with that of open source — for basic software services — is only a matter of time.

There will always be a window for proprietary products, in my view. But open source, over time, will shrink that window. The copyright model of absolute protection for 100 years is being replaced, through open source, with a model more akin to that of patents.

And that's a good thing.

Hospitals are the toughest nut to crack in the IT business.

The systems they need are complex. They must be highly-networked, and they have to deal with tons of regulations.

A hospital of 75-200 beds will pay as much as $18-20 million for a new IT system. Many do without, or use old systems bought over time. Wonder why you have to fill out a paper form each time you see a new doctor, and why every doctor's front-office is filled with file folders? That's why.

For nearly 25 years the Veterans Administration has been fighting this paperwork battle. And it's been winning with a home-grown system called VistA. A few years ago Scott Shreeve, a former ER doctor, decided to build an open source business around it under the name Medsphere, with his brother Steve. (Kenneth Kizer, a former VA official, is CEO.) Medsphere's version is based on a Sourceforge project called OpenVista, built on a Linux stack.

"The source code is in the public domain, available under the Freedom of Information Act (FOIA), Shreeve explained. "It's not an open source license, it's public domain - that's a major distinction. It's the freest thing possible.

"The VA has thrown code over the fence, but there has not been a community to catch it and expand it in a traditional open source manner. A lot of the things you are used to in open source aren't there. There's no super-strong development community, just a lot of users."

So that's what Medsphere is working with. They figure they can sell 85% of hospitals a system for $7 million, against commercial providers like Cerner, McKesson, and Meditech.

OpenVista is built around an electronic health record, which covers everything from labs and radiology to clinical information. Shreeve's goal is to build a National Health Care Information Infrastructure, based on open standards.

The first thing to be standardized, of course, has to be the license. Right now Medsphere is writing custom licenses for each client. Shreeve hopes to build a version of the Apache license, or a BSD-type license, by next year, working with the VistA Software Alliance.

I think of this in terms of hassle, filling out paperwork each time I see a new doctor. Shreeve thinks of having spent 7 years in an ER, starting from scratch with each patient because he had no data on them. How many people have died because doctors made mistakes based on incomplete medical records, he asked?

Too many.

Could there be less to RedHat's strategy than meets the eye? Might it all be, well, zen? (Or Xen?).

That's the impression I gathered talking to vice president-enterprise solutions Tim Yeaton yesterday.

The onetime Compaq executive, who joined RedHat last year, said the JBoss deal was based on cultural consistency, that the Mugshot project was about insiders following their bliss, and that everything RedHat does is based on the desires of enterprise users.

For that reason, Yeaton is not worried too much about issues like license harmonization, preferring to be just as open as RedHat sees its enterprise customers wanting it to be. "We have the most open license for each technology. There are some things we build in that may be under other licenses, and it’s not our place to choose how open source projects license."

As to where RedHat is going to be physically (JBoss is based in Atlanta, RedHat itself in Raleigh) the answer was, wherever you want to be. Yeaton himself is based in Massachusetts, where many of the company's top developers (including the folks behind Mugshot) prefer to live. "We have a lot of engineers working out of their houses." (They also just opened offices in Brazil and Argentina.)

Yeaton is personally working hardest for enterprise customer loyalty, focusing on things like certification and virtualization (that's where the Xen comes in), with top-line growth being more important than margins. This is the way it is in fast-growing markets, which open source certainly is.

"All our projects are at the stage where they are ready to build community," he added, so expect to see more moves in that direction soon.

Is this the way a company should be? Or is this the way an open source company has to be?

When RedHat announced it would purchase JBoss, most of the commentary revolved around how JBoss might be absorbed.

Turns out the deal was far more meaningful for RedHat. It represented a major strategic turn toward applications.

This was made clear today with the launch of Mugshot, an effort to create the capabilities of MySpace and build an open source platform around them.

RedHat was an operating system company, based in Linux. JBoss is all about Java middleware. Mugshot will build commercial applications. Everything created within all these units will live in the Web, with business models built around Web services.

It's a pretty coherent whole. And it's all open source.

Now, the question to be answered is — will it work?

Many people fighting the net neutrality are having trouble with the issue.

The name is easily corrupted, and difficult to understand. This allows those opposed to net neutrality to create groups called, for instance, Hands off the Internet.

The issue is maintaining the status quo, in which you can reach any Web site you wish, without discrimination. But that's tough to explain.

What I've suggested is free speech, freely heard.

An end to net neutrality does not threaten free speech. What it threatens is erect a barrier to entry, between your high-bandwidth thoughts and those who might "read" them.

For 100 years we became accustomed to such barriers. Only those who owned newspapers or who got the attention of them could be heard in newspapers. Radio and TV stations were licensed, and they raised barriers to entry still further.

What the Internet does, economically, is to eliminate the barrier to my being heard. I can now be heard as easily as David Broder. We're both using the same medium.

It's when we get beyond words, to the equivalent of radio and TV, that the phone companies want the barriers to go up. Frankly, those who control our radio, TV and film media want those barriers raised as well.

And there is no First Amendment argument in their way. You have the freedom to speak, and publish. The Constitution did not contemplate equal rights for the distribution of all words.

But if Jefferson were blogging today, maybe it would.

In a fascinating post on the aftermath of the Eolas decision, our own David Berlind note that some patent holders are becoming frineds of open source.

After losing to Eolas, he writes, Microsoft was forced to drop some functionality from Internet Explorer, but Firefox was not.

Why would a company that depends on Intellectual Property rights for its existance support open source?

1. Open source provides contracts that can protect IP rights as software spreads.
2. Open source embraces a subscription business model.
3. Many patented ideas need programming help in getting to market, and open source can provide that help.

One example does not make a trend, but the concept is intriguing.

Open source is a business model. It's different from FOSS (Free and Open Source Software), although as I've noted the licenses create a continuum between proprietary and FOSS which generally leads toward FOSS.

Owners of intellectual property would do well to examine these license terms carefully. They may be pleasantly surprised by what they read. And they may also be surprised at the reception, and help, they get in using open source as a way to reach their market.

Essential to the open source argument is the idea that the basic infrastructure of the information age is just that — infrastructure — and the public interest demands it be treated as such.

This argument is generally accepted in the software arena. It is at the heart of the Internet, which is why the net neutrality battle is so hard-fought. And it is making inroads in radio, where the 802.11 frequencies are among the most heavily used in the electromagnetic spectrum.

What stands in the way of all this are the Bells. They insist that the phone lines built under regulated monopoly are "theirs," that no one else (OK, maybe a cable franchise) should be providing that service, and that they should be allowed to use their monopoly power for their own private enrichment.

Into this argument steps Bob Frankston. The Visicalc co-founder has written a satire, in the tradition of Jonathan Swift's A Modest Proposal, called Paying by the Stroll.

In this essay the part of the Bells in our present debate is played by a collection of sidewalk owners, or as they prefer to be called Transport Service Providers. You pay them for the privilege of walking, and they decide where you can go.

It's great amusement, although he's no Jonathan Swift (who is?). Still, for the present argument, it'll do.

For Memorial Day I thought it would be fun to reflect on the wide variety of successful open source business models that have appeared just since I began blogging on this subject about a year and a half ago:

• The IBM Model — All in one. You don't even have to know it's open source. You tell them what your system has to do, they make it happen, you pay the bill.
• The Covalent Model — This might also be called the project model. Make yourself an expert in a specific open source package, in this case Apache and related projects.
• The JBoss Model — Bring together superstars in a particular area and sell their pooled knowledge.
• The Red Hat Model — An open source variation on the Microsoft model. Build a basic stack, and then build on it.
• The Ingres Model — Spin-out a failed commercial product into an open source start-up, as Computer Associates did with Ingres.
• The NewCo Model — There are lots of these, companies born of new software which they chose to make open source in one way or another.
• The Me Model — This is how a lot of real-world open source support is getting done, individuals with expertise in a particular package gathering a collection of clients around them.

(The illustration is the business model of Social Ventures Australia.)

The point today is this is not a complete list by any means. There are many ways to create an open source business, or just an open source business practice. It's not like it was in the old proprietary world, everything built around a package, get big or get out. You can be any size you wish.

Maybe you have a different business model, one not mentioned here. What is it? I'd love to hear it. Great bar-b-q conversation. Or great barbeque conversation, if you prefer. Just BYOB — bring your own bandwidth.

Contrary to our fine story today, I don't think the European Community actually killed software patents today.

I think they were already dead. (That's the tombstone of my mom's family in Rhode Island.)

Over the last several years the industry has been going through a real-world cost-benefit analsyis on software patents, thanks to open source.

Patents failed the test.

Companies have had to measure the cost of gaining and defending patents, along with the revenue derived from them, against the value gained after donating those patents to the open source community.

The results seem pretty clear. Momentum for open source continues. The fear of lawyers with patent rights against the open source movement has gradually subsided.

If open source had a business case at all, this result was inevitable. Sharing basic code and building on top of it results in higher levels of functionality and reliability than using lawyers to protect your code pyramid.

The idea never made sense to me in the first place. What are you protecting, the math or what it does? If you're protecting math — and software algorithms are just math — it's a silly gold rush. If you're protecting what the software does, then you're not protecting a mousetrap, but the idea of catching mice, which is equally silly.

The patent system, by its nature, is designed to lead toward new invention. You must publish what you patent. This allows inventors to see if there are other ways to do the same thing, and patent those methods. But once you've published source code, it's out there and policing plagiarism, especially as software grows more complex, becomes a needle-in-the-haystack problem.

Meanwhile the software world goes on.

So rest in peace, software patents. Now let's have other dreams, and better. Let's build something new on what we've got.

One great achievement of the term "open source" is to create a licensing continuum. (Eric Raymond, left, author of "The Cathedral and the Bazaar."

The GPL is a great idea, but it is the deep end of the pool. The idea of total freedom, and the responsibility to give-back in maintaining that freedom, is an awful big step for corporations to take, steeped as most are in the idea of "intellectual property."

Open source lets companies go just as far down this road as they wish. They can call it "blended source," or "mixed source," they can support a wide variety of BSD licenses, they are free to experiment, to set their own strategies.

What most discover, over time, is that they get only as much as they give. Blended source companies get less from their communities than BSD companies, which in turn get less help than those using the GPL.

Ad big companies seek to reduce their expenses for maintaining and growing software, they step out futher along the continuum. This is often done through donating code, or supporting projects, rather than through changing license terms. As they see contributions come back to them they become increasingly generous.

Getting to the GPL, in other words, is a long journey. But thanks to the term open source, there are now 1,000 steps along the way. And without it, most firms would not take that first step.

One of the biggest problems for those who want open source databasing is scaling.

And now it may be solved.

Once you start storing terabytes, and want to do deep business intelligence analysis on that data, a simple mySQL installation on a Dell box is not going to cut it. Proprietary companies have long taken advantage of this, warning firms which hoped to grow fast that they would hit a wall without the right solution.

Greenplum says it has the answer. It's a proprietary layer on an open source PostgreSQL database, enabling true enterprise-class database support with an open source base.

Greenplum president Scott Yara is not blowing smoke here. He has a reference client for this solution, Frontier Airlines. He estimates Frontier is gaining $4-8 million in extra revenue per quarter from this.

"They take data from their core registration system and analyze it to optimize routes and fares." But they couldn't upgrade to a Teradata data warehouse, as competitors had, once they got to 800 million records. So CIO Bob RathRapp looked to Greenplum's "share nothing" system.

Greenplum partnered with a business analytics consultant, DaxbyDaxpy, to come up with something that could run on a collection of low-cost servers. It wasn't cheap, but the underlying technology was still PostgreSQL.

All this works because Greenplum works directly with the Bizgres community, which applies PostgreSQL to BI problems. "We go through great effort to make sure we in the commercial version have full compatibility with PostgreSQL. So when they get people trained they are using Postgre. When people buy into us they buy into Postgre."

So, is the open source database problem solved? Yara thinks so. "We're on a real tipping point," he says.

Maybe. Just don't expect to see Michael Hayden wearing a Tux t-shirt any time soon.

Critics call them vultures, or buzzards. I prefer the term gleaners.

They're the people or companies that sweep the last crumbs of profit from a market or industry. Day traders, hedge fund managers, private equity funds — they all provide liquidity, lubricate the economy, and keep the system efficient.

Signate is using open source to glean what profit remains from the telephone equipment market. They sell a virtual PBX, based on open source tools, which small telephone companies or landlords or private companies can use to provide business-class voice service to hundreds or thousands of desks.

"The telecom business is into its last residue of proprietary applications," CEO William Boehlke told me. "It's where companies like DEC went to die. What used to be a quarter million dollar piece of gear now sells for $75,000, and we can build them for $25,000."

While a company like Fonality will build the box and sell it to a business office, Boehlke sees this as still being a service provider business. "Not only do we need a richer feature set than Fonality but we have to give customers the ability to provide E911 and 411 and meet the CALEA regulations," he explained.

"People bought PBXs so they wouldn’t pay to call a colleague. We don’t have that problem. We think the PBX wil disappear, and will become part of what you pay for basic service. Only large organizations will keep telephony in-house."

Some 20 customers are looking at the early code of the current product, there are 100 firms on his customer list right now, but Boehlke thinks there may only be a universe of less than 1,000 firms he can sell to.

That's OK. It's all profit. And with open source, Signate figures on gleaning a little more of it.

The Free Software Foundation brought its campaign against Digital Rights Management to Seattle this morning, in the form of a "flash protest."

The idea is that DRM is Defective by Design,  that it cripples computing solely for the benefit of a few, and that it needs to be gotten-rid of.

As one activist wrote me:

The fight against DRM is the fight for the right of private ownership of computers, and for the practical power to control the ones we own.

This has always been the political split between the Free Software movement and the Open Source software movement. The former sees freedom as an obligation, something that must be given to everyone. The latter see freedom as an option, a grant to others under strict license control.

The former view holds DRM in contempt. The latter is either agnostic on DRM or actively embraces it. (This is separate from the question of whether DRM creates monopolies.)

So, are the two issues connected? Should they be? Or are those demonstrating in Seattle today a lunatic fringe?

Jitterbit, which makes software tieing applications together, is going open source under the Mozilla public license, slightly modified to protect the trademark.

Two versions are being offered, a free community edition and a $9,995/server/year (for support) professional edition.