Defining Lame

So, a good friend from Stanford, Matt Harrison (aka "The Real Matt") challenged me on my contention that most open source startups today are somewhat lame. That might be because Matt works for Spikesource, one of the more recent open source startups. Regardless, I did say **most**, and wasn't necessarily referring to Spikesource. I don't think I yet know enough about what they're really doing to be able to categorize them one way or another, though I think highly of Matt, Murugan, and Kim, and so am willing to give them the benefit of a doubt.

Still, I see far too many startups emerging whose understanding of open source doesn't get far beyond the ability to cite aphorisms from Eric Raymond's The Cathedral and the Bazaar. While it's fair to separate open source development from businesses that leverage it, I think the best businesses will be those that significantly interact with open source development communities. In many cases, startups will emerge from open source communities (MySQL, JBoss, Funambol, Tenable, etc.). In others, they will replicate the energy of such a community (SourceBeat is an interesting example - I hope it succeeds).

Regardless, one's value needs to be clear. I don't see how a separate company can leverage open source without being involved with the relevant communities. Red Hat and Novell are examples of this - ostensibly, they exist separate from the Linux community. In practice, however, they employ significant percentages of Linux kernel developers. This may not give them control over Linux (it doesn't), but it does put both in privileged positions to support Linux.

So, to The Real Matt's question, while I don't think Spikesource is lame, I do think they need to clearly define how they're going to provide better service/certification/support than the communities directly involved. The LAMP stack, for example, is already supported by Red Hat and Novell (in conjunction with the Apache community, MySQL, and Perl/Pyton/PHP (Zend)). It may be that Spikesource (and SourceLabs + OpenLogic, as well as newcomer Bridge53) can effectively support various combinations of open source communities without being integral parts of them, but I wouldn't bet my money on that proposition.

[Disclaimer: Again, I don't know the "secret sauce" driving Spikesource and SourceLabs, so I'm not in a good position to categorically state that they're not planning to work closely with these communities. FWIW.]

Where are all the good open source startups?

I've been watching the open source startup space very closely - probably closer than any other person on this planet - over the past year. After talking with just about every startup that has even thought about existing, I have just one word to describe what I've found:

Lame.

There simply aren't many good open source startups out there. MySQL is a major diamond in the rough, as was Red Hat before it. But they're the anomalies.

I don't think this means that there won't be more open source success stories. Rather, I think it means that we're still very early into open source's maturity. It still has some growing pains to endure before it becomes a hard-core capitalist tool.

Maybe that's a good thing?

Object-oriented Startups

I'm not sure who recently suggested this phrase to me, but it seems accurate. Startups can rent-a-CFO; they can outsource development (or manufacturing) to India or Asia; and now they get a tremendous amount of free, foundational software in open source. This is a great thing (for startups), as it dramatically lowers the barriers to entry. I met with SugarCRM again today, and John Roberts related to me that his development team took three months to code the first iteration of their CRM product, and then rolled it onto SourceForge. I kept thinking, "InstaCompany."

The only problem I see in this trend is that as humans, we tend to be very good at taking and somewhat weak at giving. Open source only works to the extent that developers give back some percentage of the value they derive from the freely available code. To some extent, open source licenses require this. However, for web-based businesses, this same requirement does not apply. If I'm delivering a service over the Internet, it strikes me as very plausible that a court would not find me in violation of the GPL for keeping my modifications private.

If true, this means that we are generating a new generation of open source free riders. This is unfortunate, given how much exceptional code would be openly, freely available if we could find ways to induce these developers to return value to the code base they borrowed. Open source's biggest problem right now, to my mind, is that it's a fairly limited community. Despite all of the hype about a massive pool of altruistic open source developers, the truth is that there is a finite number, and that number is relatively small. If we could get all those who benefit from open source to give back to the commons, however, the pool of open source developers would swell to truly gigantic proportions. In the process, even better software would be developed (and provide a foundation for the next generation of "even better software").

I'm not into forced sharing, but I do wish we wouldn't pilfer from the commons.

What Price, Integrity?

Last year I co-founded the Open Source Business Conference. The conference was a major success but, as I reflect, not as successful as it should have been, and as it was meant to have been. We raised gobs of sponsorship dollars, had high-profile speakers, and a 96% approval rating within our attendee base. What’s not to like?

As it turns out, the sponsors.

As my friend (and critic) Phil Windley opined in his blog (http://www.windley.com/2004/03/17.html), whereas much of the content was exceptional (Lessig, Clayton Christensen, etc.), a significant amount felt like it had been bought and sold. And it had. Sponsors demanded (and got) speaking slots based on their willingness to pay. While we tried to hold to editorial values, once the money was in our bank it changed everything.

For the worse. It soured and corrupted the message of the conference, which was (and is) that there are exciting new business strategies for vendors and CIOs enabled by open source, and that the Open Source Business Conference is the place to find them. People came for insight and experience – in some cases, they only got product demos.

[Quick note: Stay with me here. This has much less to do with conferences – which is not a very interesting topic – and more to do with how the technology industry sells itself.]

I’m not sure what motivates hardware and software vendors to do this. It strikes me as a Neanderthal way of viewing marketing. Think about it: when was the last time you enjoyed a TV commercial? It happens, right? But does that enjoyment derive from humor or some emotion surrounding the product, or does it derive from a product fact sheet posted on the screen? I’d bet my underwater stock options that it’s the former.

As human beings, we distrust product pitches. We understand their bias, and resist. Why, then, do vendors persist in trying to force-feed their inane product pitches to buyers? And in such bald-faced ways as boring audiences into a hypnotic trance?

I like Matthew Szulik for this reason, among others. When Matthew keynotes, he keynotes for the open source movement (and, implicitly, for Red Hat as a key figure within that movement). After Linuxworld 2005 (San Francisco), I heard complaints from other vendors: “What was he trying to sell, talking all that stuff about patents?” They missed the point. He was selling a general feeling, not directly Red Hat AS. Assuming buyers bought into that feeling, they would likely look to the person/company that helped them feel that emotion.

Chris Stone, then Novell’s Vice Chairman, did this at OSBC2004, as well. He talked strategy, rather than products. He highlighted a key difference between the strategies of Novell and Red Hat: Novell believes customers benefit from a ‘Both Source’ IT strategy that pragmatically adopts the best of the open source and proprietary worlds. Less religion, more practicality. It was an honest message, and was well received. The media, experts at sniffing out the chaff from the wheat, gave him great exposure, as they did Szulik’s Linuxworld keynote.

Interestingly, many of the other OSBC keynotes were not heard beyond the Grand Ballroom at the Westin St. Francis, where the conference was held. Why? Because nobody beyond the presenter cared about the message.

Of course, this principle holds true beyond IT. I’m increasingly concerned by how television, movies, magazines, and other media are increasingly commercialized. It is impossible (or, at best, extremely difficult) for the creator of a piece of art to remain independent of those funding it, just as it was exceptionally difficult (and, in the end, impossible) for me to deliver objective insight through my speaking faculty at OSBC2004, when many were merely mouthpieces for software or hardware vendors. One may dislike the bias of Hollywood and the media, but at least traditionally it was a political bias, rather than a product bias.

This may seem like an insignificant issue, but I disagree. It is important that we have sources of objective information. Good decisions are made based on good information. Bad decisions are based on corrupted information (or foolishness, but there’s nothing I can do about that).

We started our fund-raising for OSBC2005 with the expectation that we could return to a focus on content while still raising gobs of money from sponsors. Soon into the process, however, we got tired of hearing things like:

“Will I be able to send product information to conference attendees?”
“If I pay more can I have a closed-door session with the CIOs present?”

And so on. It actually started to make me physically ill, as I thought of what we were going to be forced to inflict on our attendees, all in the name of raising a few dollars in sponsorships.

So, we’re not.

This year, the Open Source Business Conference has made a conscious decision to seek less sponsorship money. Our goal is to deliver cutting-edge content on open source trends and how they affect the enterprise, both CIOs and vendors to those CIOs.

Our keynotes have been solicited based on their intelligence and insight, and not on their ability to pay. That’s why we have Jonathan Schwartz (President and COO, Sun Microsystems), Marten Mickos (CEO, MySQL), Larry Lessig (Author, Free Culture and professor at Stanford Law School), Geoffrey Moore (Author, Crossing the Chasm), and others keynoting. (Others will be announced soon, including some high-profile speakers that will surprise many because of their seniority in the industry, among other reasons.)

There will be no vendor pitches from the OSBC pulpit in 2005. We’re reclaiming the conference from the sponsors. It will be more like DEMO or D or OSCON, and less like Linuxworld (vendor exhibition) or the various CIO conferences (which are really just glorified vendor pitches).

In closing, some advice to hardware and software vendors who are hoping to play in the open source world. Open source is all about community, and community must be created. It's one thing to ride on the coattails of Linux or Apache or one of the other big projects, but it's quite another to actually foster and grow one's own open source communities (which is where the money is/will be). That sort of community development requires that a vendor position itself as having a clue about open source, and share that 'clue.'

Such positioning requires actual intelligence and novel strategy. Buyers aren’t stupid. They can differentiate between a canned infomercial and true thought leadership (and understand that thought leadership will lead to product leadership in this fluctuating world of IT).

This positioning also must be done vis-à-vis one’s partners. Red Hat has lead in the Linux market because would-be partners have sought them out as the thought leaders in open source. That lead is shrinking now as Novell has offered a strong, and different, open source strategy.

In the open source world, marketing must be more thoughtful, because the open source community can separate the wheat from the chaff with a slight twitch of Slashdot. That means selling one’s vision, rather than simply one’s products.

Bringing this all full-circle, it means sponsoring for a larger purpose than for blistering a conference’s attendee base with one’s product literature. No one wants it. Really.

PalmSource, China MobileSoft

PalmSource's announcement has not been fully appreciated by the open source crowd. That's understandable, as many in that community believe that open source is the sole value creator. On the other planet, companies like Microsoft believe that proprietary products are largely the only way to go.

As PalmSource shows, however, the truth is in the middle. Linux is a great embedded platform...but it comes with practically zero apps. Palm OS has a wealth of apps (and those who say PocketPC has surpassed it have evidently not seen the devices making a splash on the smartphone scene - it's the only market that truly matters going forward, in phones). The marriage of Palm OS and Linux is one made in heaven.

OSDN Article: Open Distribution vs. Open Source

I published this a few months ago on Newsforge, but in case you haven't read it....

The article traces the near-term value of open source to startups. Distribution, rather than development. (And, depending on the nature of the startup, it might equally be about free building blocks, as Friendster has used open source.)


The Open Source Capitalist
By Matt Asay
For OSDN (Newsforge or ITMJ)

The White Rabbit has beckoned us down the wrong rabbit hole. Much has been made about the open source revolution, and with good reason. The open source development model produces superior software.

But the promise of open source lies not in open source, but rather in open distribution. (The fact is, very few customers ever look at source code, and even fewer modify it. In most cases, modification violates vendor support contracts, anyway, so customers have little incentive to poking around in the source.) This March’s Open Source Business Conference (www.osbc.com) and private conversations with key people at MySQL, JBoss, and other successful open source startups has me convinced that open source is about much more than software development. It’s about business development.

In other words, what we thought was a software development methodology has perhaps even more importance as a business strategy that leverages its minimal sales, marketing, and distribution costs. Perfectly following Clayton Christensen’s arguments in The Innovator’s Dilemma and The Innovator’s Solution, open source proves to be an exceptional business model for upstart, low-cost competitors to take on the incumbent vendors.

We are sitting on the most exciting IT business model capitalism has ever seen, all thanks to the GPL. It’s a model that will take software commoditization – started by Microsoft two decades ago – to the next level, pillaging all those who fail to capitalize on this new model. In the process, open source will dramatically expand the software market by making IT affordable to hitherto excluded would-be buyers, much like the lowered cost of cell phones and other technologies ended up massively inflating their markets.

Richard Stallman and Adam Smith – bosom buddies?


THE NEW INTELLECTUAL PROPERTY

I used to think that open source was relegated to the infrastructure layer of the software stack. Indeed, this continues to be where it thrives. But as entrepreneurs begin to grasp the possibilities open source engenders, no area of the software stack will be immune from open source commoditization.

Let me explain.

Intellectual property (IP) law (at least, here in the US) was designed to ensure the ability to profit from one’s creations. I write a piece of software; I copyright it; I sell it (assuming it’s a useful piece of software and I’ve adequately marketed it so that people know about it). Simple.

This has been the software industry’s dominant model for decades, and has created a few mammoth software companies who have successfully exploited their IP to generate billions of dollars in revenues. In this model, exclusion (i.e., the ability to keep competitors or customers from copying one’s code and replicating it to others) equals profits.

In the open source world, IP continues to play a role. But it’s a different kind of IP. Dubbed ‘copyleft,’ what it essentially means is that by freely giving away my crown jewels – the source code – under the GNU General Public License (GPL) or similarly restrictive license, I benefit my customers while leaving my competitors no better off. Yes, they can take my code but, no, they cannot borrow it and drop it into their products (without returning the favor by giving me their code). In practice, few to none of today’s successful software companies are going to take this risk…at least, not until it’s too late.

The GPL, then, offers a way to exercise an incredible amount of control over one’s competitors. By open sourcing my code under the GPL (and then competing vigorously for mindshare so that customers desire to use my product), I force down my competitor’s prices and significantly out-maneuver them by skipping over Procurement and every other barrier to getting my product in pilots (however informal) within my would-be buyers’ companies.

Open source also enables me to maximize my market penetration at minimal cost. MySQL has 5 million downloads today. 5000 of those would-be customers have returned to buy a support contract/license from MySQL. MySQL’s sales and marketing costs go way down because they have a growing universe of pre-qualified buyers coming to them. Would-be buyers are just a download away, which download need not go through Procurement’s cumbersome bureaucracy. Distribution costs go down, as well – it costs next to nothing to set up an FTP server.

And, depending on whether the open source company is the “owner” of a project (MySQL, JBoss, SugarCRM) or an active contributor (Novell, Gluecode, Specifix), they may or may not be able to lower development costs by borrowing a rich foundation of code contributed by the community. Even if the vendor employs all of the developers who contribute to a project (MySQL), they still benefit from a global QA effort, wherein users of the code contribute back bug reports/fixes.

Sounds OK, you say, but what is to stop the competitor from simply forking my code and selling their own version of AsaySoft. Nothing, I suppose, except human nature, and that proves to be a very potent retention tool. I’ll explain.

In this new world of open source, reputation capital means as much or more than traditional intellectual property/capital. If I employ the developers on a given project, I have a measure of control over the direction that open source project will take. But even more importantly, the more developers I employ who work on, say, the PostgreSQL database project, the more likely it is that would-be customers will trust me to be able to support it. Once a company is thought of as the default vendor for support of a given project, the harder it becomes to dislodge that vendor. This jibes perfectly with other commodity businesses where brand, cost (how cheaply I produce something governs how cheaply I can sell it, and low prices sell, as Wal-Mart’s $260 billion in revenues evince.), and service provide the only lock-in.

And it is lock-in, indeed – the good kind.


CUSTOMERS WIN IN THE OPEN SOURCE WORLD

We are rapidly nearing the point where mainstream software will be commoditized by open source, especially as an increasing array of startups leverage the open source business model. This will tend to push all margin-generating value to the services that support hardware and software, with brand and lower-cost models augmenting that margin.

In such a world, customers benefit as vendors must focus on solving business problems rather than creating corrals (lock-in) to keep customers from escaping shoddy software and service. The days of installing monolithic software packages that create more problems than they solve – where the services required to make a piece of software actually function as advertised cost 4-10x what the software itself costs - are nearly over. This will leave more money for CIOs to do what they’ve always wished they could do – spend for true value-adding services that tailor IT to meet their specific business demands.

But it’s not completely a rout by the customers. Vendors benefit, too, because they will be forced to intimately understand and solve their customers’ requirements – the ultimate lock-in. If I, as Novell, solve Customer X’s problems more efficiently than my competitor, I can virtually guarantee that Customer X will buy from me because they want to…over and over and over again.

Copyright and patent are weighty protections, but they put the vendor in an adversarial relationship with the customer. Such traditional intellectual property tools hurt users as much (or more) than competitors. Open source allows me to lay waste to my competitors’ profit margins while simultaneously blessing my customers with increased IT flexibility and a more finely-tailored approach to solving their business problems.

Now that is innovative.

Big Bang Event: My First Open Source Paper

I wrote this paper while at Stanford Law School, working under Larry Lessig. Most of it is still true. (Click the title above to read the paper. It's in .pdf.)

Mensa Article

A few months back, I spoke at the London Chapter of the Mensa Society. Afterwards, they asked me to encapsulate my presentation into an article they could publish in Mensa Magazine. I did so, and it was published earlier this year. (For those looking for it, it's the one with the "Mensa babe" on the cover - "Beauty and Brains," I think it was headlined. I was tucked away after the pin-up. :-)

A basic article, but it lays out (briefly) my thoughts on how open source replicates the scientific method (Popper's, not Baconian induction).


A New Choice in Information Technology:
The Rise of Threatened Fall of Open Source Software


Introduction
Something is happening to your local software vendor. Something big. Something that threatens to topple Microsoft from its hitherto unquestioned reign of monopolistic dominance.

Something called Linux. Something called open source.

Open source is not a thing. It is not something you buy at PC World or Dixons. Nor is it something that your company’s system administrators install on your computer. It’s not an American thing, or a British thing, or, really, an “anyone thing.” Rather, open source is a globally recognized and practiced software development methodology, and one that has Microsoft scared out of its US$56 billion bank balance.

Why? Why should an alternative development methodology, one that closely follows scientific methodology, frighten Mighty Bill Gates?

Because open source shifts control toward buyer, away from the exclusive control of the vendor. Open source brings choice and accountability to an industry that has long promoted expensive “solutions” to business problems, solutions that have proliferated more problems than they have solved. Open source’s most important contribution to information technology (IT), then, is the very thing that makes it most unpalatable to the world’s largest software vendor:

Open source creates choice.


How Open Source Works
Traditionally, software has been created by software and hardware companies to fill customer needs. For example, Novell determines that a market exists for collaboration software (email, calendaring, etc.) Novell hires engineers, engineering project managers, marketing directors, etc. to scope out what the product should look like (functionality), architect it, and then build, market, and sell it.

Open source functions in much the same manner, except that there are no marketing and sales people, and the engineers who collaborate to create the software are employed by disparate corporations, or are self-employed. There is no CEO determining arbitrary project roadmaps; rather, the open source project leader works with that project’s community of developers to determine features, roadmap, and everything else that will go into creating the final product.

Because the developers are geographically dispersed, they communicate through email, instant messaging, and other Internet tools. Few ever physically meet each other. As such, they must write modular code that can be “snapped together” like Lego blocks to create a final product.

In the open source world, no one gets paid for her work (unless her ‘day job’ employer chooses to allow them to write free software), so motivations for contribution vary widely. Importantly, then, such developers tend to only write code that which interests them. This is both good and bad. It is good insofar as this self-selection yields better code, as engineers choose to develop software that interests them and that they have aptitude for, rather than what her corporation tells them they must. It is bad inasmuch as the process of developers writing for other developers leaves out a lot of great software that needs to be written. Hence, much of the software created through the open source methodology tends to be ‘geeky,” highly complex programming like that involved in creating an operating system, database, or other infrastructure software. The biggest exception to this is OpenOffice, a competitor to Microsoft Office, but OpenOffice developers are overwhelmingly employed by Novell and Sun Microsystems, and get paid to contribute.

Normally, a decentralized approach to software development would likely fail, but open source works because of the licensing regime fueling it. Open source licenses, with the GNU General Public License (the license that governs Linux) prominent among them, generally require:
• The right to access and modify a given piece of software’s source code, or the human-readable language that tells a computer what to do;
• The right to distribute this modified source code (“derivatives”); and
• A requirement that distributed, derivative works be licensed under the same license under which they were received.
In this way, open source licensing promotes enforced sharing. You can only receive if you agree to give.


Science and Open Source
Open source works because it functions much like science. Contrarily, software written in the traditional model proves flawed and buggy precisely because it fails to conform to accepted scientific measures. By science I do not refer to Sir Francis Bacon’s inductive Scientific Method. In this model, observation gives way to a hypothesis as to why something happens; that hypothesis is tested until it yields a theory; this theory is then elevated to a law (“Truth” with a capital “T”) when it shows itself adequately resilient over time. This is not, to my mind, how science actually operates. In fact, science is not so much concerned with proving things true. Rather, its real aim is to prove things false, as Karl Popper taught. Scientists develop “conjectures” about how the world works, and then enter into a process of “refutations” by which they attempt to find holes in their analyses.

Just as science proceeds through a series of “conjectures and refutations” – i.e., as scientists try to test their observations into oblivion, and provisionally accept certain observations/theories as not-yet false – so, too, does open source work. Open source is not concerned with creating Perfect Software through a brilliant design. Rather, open source aims to discover robust software through a robust peer review process.

This differs dramatically from traditional, corporate software development. Unlike a corporation, which may prefer to upgrade software and fix bugs in its products when convenient or profitable for them, the open source community is only concerned with making software better. This means fixing bugs upon discovery. This is why open source is so critical to good information technology (IT): just as Popper analogized from science to society, arguing that society required open, “falsificationist” debate in order to flourish beyond dangerous dogma, so, too, does good software require that it be written outside the auspices of any one company’s exclusive control. Good code, like good government, depends on the ability to perceive and correct inefficiencies and imperfections in the fabric of that code (or society), which ability is lessened by centralized control.


The Closing of Open Source?
Given open source’s rise to prominence, it is not surprising that corporations increasingly seek to influence and control it, and hence taint open source’s chief benefit (its openness). But we should resist this impulse. At this still-nascent stage of open source’s development, it is critical that open source be nurtured and protected.

Corporations have tried various means of controlling open source, but none so extensively as what I call the “Infiltration Model.” Because open source functions through a community development model, the more members of that community that a corporation “owns” (i.e., employs), the greater that corporation’s influence in where a given open source project will go. IBM, Red Hat, Novell (which recently acquired Ximian and SuSE, two prominent open source companies), and Hewlett-Packard employ a huge majority of developers of the Linux kernel, which offers them a way to guide Linux’s development in a way favorable to them.

True enough, all people and corporations have the ability to compete in exercising influence in open source communities, arguably canceling each other’s influence out in this free market exchange. Even if this is true, we should not be so naïve as to believe this will not have an effect on Linux. At the foundational layers of the software stack, the wider the community involved in developing them, the better. The more such development is centralized in one or a few companies, we begin to reassemble the Microsoft world. The platform should be open. Higher up the software stack, at the application layer, let companies compete on offering true innovation. There, “propertized” innovation is necessary and proper. But not at the foundation. The foundation should be free.

Intellectual property, then, long the stick by which corporations have fended off competitors and justified high prices to customers, is giving way to a new form of property ushered in by open source. This new “property” is reputation. Reputation capital in the open source community will come to mean as much or more than control of intellectual property, which is why corporations are flocking to build or buy reputation in open source communities, hoping to exert control over their product roadmaps and the customers who buy into them. We should resist.


Conclusion
Innovation flourishes best where it is controlled least. The Internet is an obvious example of this, as is open source software. Open source has created the software that powers most emails that we send (Sendmail), most websites that we visit (Apache), and an increasing number of the servers that power our corporations (Linux). Whether it will continue to do so largely depends on whether corporations can resist the temptation to exercise undue influence in open source communities. Let us hope that they will. Even better, if we have the ability, let us involve ourselves in open source development, providing a check on corporate control and an assurance that open source will remain open.

Square Pegs, Round Holes

I've talked with just about every open source startup this year, and have one conclusion: we still don't get open source. To the extent that open source, being a different development model, demands a different revenue/business model, we're very much stuck in the 80's with our poofy hair and Def Leppard.

There are some, like MySQL, who have maximized open source's potential to lower distribution (plus sales and marketing) costs, and that's great. But the new startups that aim to certify and deliver open source stacks strike me as a poor way to deliver the richness of open source. If the problem is vendor lock-in and IT's inflexibility, then why is the solution to cobble together open components and attempt to close them through a certified stack?


I understand the business reasons behind doing this. I agree that there is a need to reduce the complexity of a myriad of open source projects into consistent, consumable "stacks." But I still wonder if vendors are attempting to cram buyers into the new world of open source IT - a world for which buyers are clearly not ready. How do I know this? Because buyers are still looking for vendors to remove all that is fine and good (besides low cost) from open source before giving it to them.

It may be necessary to wait for a spell, for buyers to catch up with vendors.

The Customization Economy?

So, I was talking with Bob Gett (Founder and CEO, Optaros) the other day, and he made a comment about how open source enables the customization economy. I liked the sound of the statement, but it's only now that I'm unpacking it to figure out what it means, and whether it's true.

I suppose it means that access to source code allows one to customize one's software in a myriad of ways, and that a trail of money will follow the company that facilitates it. Maybe. At least, it goes along with the general euphoria often associated with open source. But it's not clear to me that customization = revenues.

That's the tricky part: economy. Is there a growing economy around open source? Yes, as attested by IBM's, Red Hat's, Novell's, etc. revenues related to Linux. But is it really as exciting as we routinely trumpet?

I don't know. In fact, in some ways, I'm wondering if we aren't overly focused on open source as an end, rather than as a means to lots of exceptional ends. I'm also not convinced that buyers care about customization. Sure, everyone likes to have things their way, but only at a reasonable price point, with a minimal amount of risk. Even then, for the wide majority of enterprise buyers, choice/customization is something to be sacrificed for convenience and cost. That's why Red Hat does so well selling a vanilla, standard version of something (Linux) which, in theory, people should be ecstatic to customize.

Customization is not an end. Productive IT is the end. Open source may be one way of getting there, but it's not the only way. In fact, it may not even be the most important way.

For example, a company with which I'm increasingly familiar, CanyonBridge, has super-cool technology that allows them to integrate disparate data sources/applications and "last-mile" deliver them to the end user through a super rich, web-based UI. An EAI tool like this could be the means of tying together open source apps (hitherto banned to tepid use because of their lack of polish) and delivering them through an interface that is superior to Windows. Have Linux + browser, will compute. The importance, however, is in the integration, not the customization.

It's nice to personalize one's ring tones, our clothing, etc. But we listen to ring tones that mesh well with the musical tastes of others, and wear clothes that fit in with those of our neighbors. In IT, we're even more conservative. No one wants to be the oddball.

Where this changes is in the builder economy. Appliance vendors and other OEMs, as well as embedded device vendors love Linux/open source because it offers them malleable substance with which to build products. Customization is king for them, because it equals competitive differentiation. IT departments don't really want this - they just want to make the trains run on time. But a Juniper Networks **must** have this customization, because they want to make their trains look better, to draw in customers.

So, I suppose, whether or not it's a customization economy depends on which market you're playing in.

With this in mind, it strikes me as funny that the enterprise gets the most attention relative to open source, when it actually does the least with it. The enterprise buys standardized versions of Linux, and new startups (Spikesource and SourceLabs - OpenLogic does something different, and more akin to the builder economy) have sprung up to take all of the customization out of other open source components, as well. It's humorous, really, when you think of how the greatest thing to happen to enterprise open source is that companies like Red Hat have come along to effectively close the source and remove choice. It's only in the embedded and OEM markets that open source still lives up to its promise of access and customization.

Side note:
I guess this means that VCs should fund Specifix, instead of scratching their heads and wondering why Kim and Erik are working so hard to enable customization. The problem, guys, is that the VCs can't think outside the enterprise, where code "wants" to be closed. (My conversation with security appliance vendors last week - e.g., SonicWall, Juniper Networks, etc. - tells me that they definitely live in your world. Now you just need to find some VCs that have familiarity with that world.)

Ugly Linux

So, in addition to my Mac, I'm now required to lug around my standard issue Linux machine. I know: as part of Novell's Linux Business Office, this shouldn't be a requirement, but a badge of honor and a privilege that I proudly bare to the world. Instead, I bear it.

The Linux desktop is somewhat overhyped at present. In the mad dash to establish Linux as the heir apparent to Microsoft's long-held monopoly, vendors (including my employer) have sometimes been too eager to tout Linux's benefits without also admitting its weaknesses as a desktop OS.

More than this, I fault us for not taking enough care to think through the novelties that web services, among other things, offer us in the way of desktop apps. Why are my apps tied to my computer at all? Why aren't they housed on a "key" or "thumbdrive" in the way Realm Systems (www.realmsys.com) does? Or why aren't they delivered over the web in the way that CanyonBridge does? Why do we persist in trying to replicate Microsoft's mistakes?

The Internet is the computer. We're finally getting to the point where this claim is as substantive as it is true. There is no compelling reason to continue to cling to the PC. Apps should be omnipresent, not tethered. And Linux should stop trying to clone the anemic Windows experience and instead adopt "etherware" or, at least, mimic the Mac.

Yes, I'm calling desktop Linux ugly. It may be your "baby," but it's only going to grow up and be worth gawking at if we leverage the benefits of open source (Linux is a network operating system, not a Microsoftish client system) and move beyond Microsoft.

Lowering "Costs" in Communication

I remember growing up in Virginia; how important proper language was. I was threatened with detention for not addressing my teacher as "Ma'am." (Later, when I moved to Casper, Wyoming, for sixth grade, I was threatened with detention for calling my teacher "Ma'am." She thought I was "sassing [her]." Different cultures, but the same requirement: speak properly.

Email, IM, and now blogging are lowering such artificial barriers to communication. If you've experimented with IRC chatting or more mainstream IM (Yahoo, AIM, MSN), you know that shorthand is the norm. It's just too cumbersome to type everything out completely when the same thoughts/feelings can be rifled out. Texting on phones is an even more extreme example of this.

Some might complain that the quality of language and communication is deprecated by the casual looseness of language in these media. I disagree. I'm an English major, a lawyer by training, and highly anal retentive when it comes to proper grammar and spelling. But I think this "loose" communication is simply a way to have casual conversations online, rather than in buttoned-up briefs, memos, and "Dear Sir or Madam" letters.

Blogging, for its part, allows would-be publishers to throw their thoughts out to a global audience, without the ritual emasculation that editors and hierarchy inevitably impose. Yes, accuracy and good sense may sometimes be lost with the oversight, but more is ultimately gained than lost. It's more important that something be said than it be said exactly right, with perfect punctuation and just the right tone.

The Internet and the computer are reinventing the casual, content-rich conversation. It's something we lost for awhile as we experimented with typewriters, word processing programs, and such. I'm sure that these free-flowing communication media will calcify over time but, for now, they're rich channels of communication.

Hope in Open Source

Tonight I had a fantastic conversation. It's not one that I can relate at this point - I'm under "moral NDA." That is, I promised not to tell. And so, I won't.

But the nature of the conversation was this. An exceptional company plans to open source, as it were, an amazing body of code. (It's not that they're actually releasing formerly proprietary code as open source, but rather that they plan to open up their code to the wonderful world of open source by making it compatible with Linux. Trust me - it's bigger news than I can convey here, until they make it public.)

They wanted my opinion and guidance on the best way to do this without running afoul of open source codes of ethics and proper dealing. My opinion? They're going to have a tremendously positive impact on Linux and the entire market for open source.

What I found so gratifying was the enthusiasm in their demeanors. They felt that they were doing something significant by sharing, and so they were/are. There's something about open source that makes for hope. I felt that nearly two years ago when Novell entered the open source waters, mostly ignorant of what it was getting itself into. But it was the promise that something positive would come of the big risk we took (potentially sacrificing the hundreds of millions of dollars in revenues from NetWare).

Open source offers hope because open source engenders new freedom - new choices. It is a way out of the IT manacle that we routinely lock ourselves into. I wish this company the best of luck. I hope to be able to play some small role in continuing to aid them in this course of action. What a great day when the news is announced.

Missing the Marketing Boat (in Open Source)

So, I co-founded the Open Source Business Conference, and we're out raising sponsorship dollars for the 2005 show. (April 5-6, 2005. www.osbc.com.) Given the success of last year's event, this should be easy.

Well, it hasn't been hard, exactly, but there have been a few frustrating moments. Like when a large IT company (which shall remain unnamed) told me that they won't sponsor the conference because it doesn't reach their target audience. I know what they meant - that OSBC is not focused on the CIO. (This is incorrect, incidentally. While the show was quite vendor-centric last year, we're significantly expanding this to target the CIO in 2005.)

But their view, even nary a CIO even walked in the door, strikes me as short-sighted in the open source world. Open source is all about community, and community must be created. It's one thing to ride on the coattails of Linux or Apache or one of the other big projects, but it's quite another to actually foster and grow one's own open source communities (which is where the money is/will be). That sort of community development requires that a vendor position itself well as having a clue about open source, and publicize that 'clue.' (This includes to one's own employees, who are constantly questioning the existence of a 'clue' within their own enterprises.) There is no better venue for doing this than OSBC.

It's also important to market oneself to partners. Red Hat's (fast dissipating) lead in Linux has largely been a function of ISV support. That support has largely been a function of branding. As Novell develops brand, partners are flocking to it. No better venue for branding than OSBC.

I could go on. It's just frustrating that companies still view marketing as rigidly as they do. Marketing is not something you do Here X to Demographic Y. Marketing happens all the time, in various venues - it slips through the cracks in the container you intended it to stay. While it's good to focus one's marketing efforts, the distinction between customer and partner is eroding in an increasingly open source world. No clear line exists anymore.

The Importance of the Uber-Consumer in IT

A few months back, I was fortunate to attend the O’Reilly Open Source Convention, the premier event for open source/Linux developers. I am not a developer, but my job is to understand this breed and to figure out how my company can leverage its direction/momentum.

Tim O’Reilly, luminary within the open source world, used his keynote to pontificate on the importance of watching the “uber-geeks” to learn where the market is going. In fact, the entire conference took on this theme, in one way or another. It seemed a bit self-absorbed at times. “Watch us, because we are the future,” seemed to be the battle cry.

I couldn’t disagree more.

It’s not a question of whether or not open source will win out in some markets. It will, SCO notwithstanding. But the point is that the emphasis is wrong. The emphasis on developers, I mean. Markets do not follow producers – they follow consumers.

Just look at Novell as a local, painful example of this. Novell has long been acknowledged as the technical leader in the network operating system market. Engineers love them, because the technology is elegant, stable, and advanced. Novell routinely wins technical accolades of every sort.

Consumers (in this case, IT departments) apparently couldn’t care less. Novell reached its peak in the early to mid-90s, and has been sliding ever since. Microsoft decimated Novell’s market share through a multi-faceted strategy, one primary component of which was to make development to its platform easy (via Visual Basic). That is, it understood the target customer base (“Sure, we’re geeks, but it would be nice to have something other than a command line now and then.”), and vigorously courted that target.

Microsoft now faces a threat from the Linux operating system, which has enjoyed early success mostly because of its low cost. For Linux to continue to make inroads, however, someone will need to make it as easy to use as Microsoft’s products.

Because that’s what customers want. Even geeky system administrators.

This doesn’t mean that technology must be “dumbed down” to meet consumer demand. It does mean, however, that to the extent that it gets too far out in front of the unwashed masses, it is worthless. Not useless, but worthless, because no one will buy it, invest in it, or care about it in any material way.

In other words, technology is not an end in itself, which is unfortunately sometimes how engineers view it. Technology is a means to an end, and that end is to service consumers (including everyone from my mom to the Pakistani government).

Even if you don’t believe this argument, believe in market sizes. According to government information, software engineers are approximately .7% of the US population. Is your technology being created to service that anemic slice of the overall pie, or is it being created to service the other 99% of humanity? It’s not hard to guess which business plan will pull in the venture capital, or notch a multitude of customer wins.

Consequently, Utah’s open and closed-source developers/IT companies alike need to get in touch with non-techies. Maybe this means focus groups, or maybe it just means talking with the marketing and sales people a bit more, to learn what customers really want. Technology is a means to an end, and not an end in itself, and should be crafted to meet the needs of everyday people (including system administrators, developers, etc.). Take a lesson from Microsoft. Let’s develop for the Technology Proletariat, whatever their occupations may be.

The Myth of Productivity

Either we’re not very creative, or we’re simply not very bright. Whatever the reason, we in high-tech (and many other industries, like the legal profession) tend to measure productivity by the number of hours we spend at work. If we “pulled an all-nighter,” we’re determined that everyone should know about it. If we’re working late, we’re sure to send an email to our boss so that they know that we were being “productive” right up until midnight.

After all, why should our productivity go unnoticed?

Of course, true productivity rarely goes unnoticed, because true productivity is tangible. It can be measured. Even our government – which we routinely harangue for incompetence - understands this. As noted on the Department of Labor’s home page, productivity is a function of output (goods and services that can be consumed) measured against labor input (hours worked). One is productive if one produces, not if one simply spends time at the office. The more one produces in a given period of time, the more productive that person is.

I used to see this all the time when I lived in Silicon Valley. My first employer was a major Japanese conglomerate which, at its peak, did ~$200 billion in annual revenues. Everyone at the company worked long hours. Well, except me. When I left at 6:00 (or whenever), my colleagues were still there, chatting around the water cooler, playing games on their computer, or simply browsing their stock portfolios. I endured the occasional dirty looks as I walked out because I knew that my performance, my productivity, routinely met or exceeded their own. It was a function of actual work, rather than hours worked.

In a big company, it’s easy to coast by on the appearance of work, rather than actual productivity. Just occasionally tell people how busy you are. They won’t believe you, but their subconscious will worship you for the size of your To Do list. You’ll get by.

However, in a startup, the wheat quickly separates from the chaff. With 40 people in a company, it becomes immediately, painfully obvious who the producers are. When I worked at Lineo, we had a sales executive who was constantly busy. Insanely busy! All over the world busy! And yet he didn’t produce. (Or, rather, I should say that because he was so busy he never produced.)

Busy-ness never was productivity. It doesn’t how many times you tell people you’re busy; it just won’t make you productive. Productivity comes from unplugging from the frenetic busy-ness that we try to hide behind. Email is busy; reading a book and stopping to jot down notes and impressions is productive. The latter activity leads to the reason you were hired in the first place: your brain. The more you unplug your brain to crank out emails and sit through meetings, the less productive you become.

Utah has a disproportionately smart and productive population. We have the 13th-highest percentage of college graduates in the nation and we are projected to have the nation's fastest-growing college enrollment until the year 2007, according to the Utah Information Technology Association (UITA). Couple that with our extremely high computer literacy rate and we’re poised for a dramatic rise in our tech industry.

But first we have to close our email client. We need to turn off the Crackberries and Treos (a sore point with me – I love checking my email at 80 MPH on I-15). We need to decline more meetings and spend more time alone or in serious conversation, working through business obstacles that plague us.

Our brains will work, if we’ll just use them. Then, all it takes is a few bold strides; diligent, productive work; the occasional big risk, and some luck and Utah will be the technology hub it is resourced to be. Then, and only then, will it be productive, Utah-style (i.e., with people leaving work “early” to spend time with their families, and returning to work after their kids fall asleep, as needed).

All for Piracy

Not airing any dirty laundry here, but I found this clip (Who Steals Movies, Episode II) very funny (and appropriate).

http://www.pro-piracy.com/media.asp

I wish the studios would give me the ability to buy movies soon after theatrical release, as the record labels have. Enough said.