On the future of open source

SearchOpenSource has a good interview with me on the future of open source. I prefer when the media whitewashes my comments, because I sound much more intelligent that way. :-)

But here I am, in all my unvarnished ignorance. Some gems:

"I might be naive on Microsoft -- there are pockets within it that are as belligerently against open source as they ever were -- but a majority of people there have a fairly nuanced view of open source. Microsoft is not always right with its perspective of software, but if the open source community tried to wall itself off and be the pizza-eating late night developers club, it could only have so much relevance."

and this on the near-term outlook for open source startups

"I finally think we are on the cusp of a mission critical, widely adopted enterprise open source application. Another great thing I see is that companies -- like LucidEra -- [are] applying to be part of the OSBC [startup] showcase[, companies that] have never been heard of before. A year ago I could say I knew every single company that was coming to the conference.

Also, more and more open source applications will be developed for every layer of the software stack by much better companies. Enterprises will think of open source as their primary way of doing business. Right now it is still nipping at the heels of traditional software, but soon companies will have to have a pretty darn good reason why they are not doing business in an open source fashion."

Varnish, please! The synopsis is this: we're seeing better management teams, better products, and better business models in open source. Ironically, as this wave rises, just being open source won't be enough. Companies will have to offer compelling technology, too.

It's a great time to be buying and selling open source software.

Mudblood open source

I talked with a prominent open source developer - Gianugo Rabellino of Apache/Cocoon fame - today, getting his feedback on the business models of Red Hat, Alfresco, MySQL, SugarCRM, JBoss, and others. He made it clear that the open source development community (if it, in fact, exists) strongly prefers pureplay open source projects, rather than "mudblood" open source companies.

No surprises there.

I then spoke to a group of architects (quite senior) from Fortune 500 companies at a recent OSDL advisory meeting. I asked them the same question: do you care if your software is open and, if so, how open is open? Almost universally, I was told that everything should be 100% open (and they then gave some good reasons as to how this would influence their decisions). They even said they'd pay for it, but I didn't believe them on that one, human nature being what it is. :-)

What wasn't clear to me in either conversation is how to quantify community involvement. None of the companies above made the pureblood cut, and yet all are making healthy revenues. Is this because each is a sell-out? Profitable sell-outs, but sell-outs nonetheless?

His suggestions?

• Hire a prominent open source developer. Google did this with Greg Stein (and Chris DiBona). Novell did it by acquiring Ximian (with Nat and Miguel as part of the package). I'm not convinced that these or other developer-hires did much for the top lines of these companies, but certainly Nat's and Miguel's involvement with Novell gave it a revived freshness (which now only real revenues can sustain).
  
  
• Actively participate in relevant open source communities. Alfresco, for one, has been doing this, but I think we could be doing better. And maybe MySQL could have benefited from more active InnoDB involvement...? :-) I know that RightNow actively contributes code to the HtDig project, as well as bug fixes to MySQL, too. Is this enough to make these companies "pure?"
  
  
• Open up all of one's source code. No "baitware," he said. So, no RHEL vs. Fedora, Community vs. Enterprise, etc. This sounds great in principle, and I admit to being biased toward openness in software, but I'm not convinced that the developer community that Red Hat lost by moving to the RHEL/Fedora split was not more than made up by the committed developers and customers it gained as a result.

And so, the question: does community matter? Or, rather, does the pureblood development community matter?

As I've written before, of the top 50 contributors to SugarCRM's development community, 95% are SugarCRM partners. Are they tapping into the pureblood open source developers? Maybe not. Are they tapping into an audience that matters and actively contributes? Yes.

Despite my apparent assurance on the matter, I'm anything but convinced by my arguments. If you have any ideas on how open "open source" should be, please ping me.

The semasiology and onomasiology of open source (r0ml redux)

At the recent SDForum executive open source event, the morning panel included a discussion of the meaning of "open source business." Tim was the first to push back on the traditional definition of an open source business, suggesting that Amazon.com, Google, and others who integrate open source into their businesses should (or, at least, could) be considered open source businesses. And Microsoft, because it contributes code, can be, too. Simon Phipps of Sun, for his part, noted that they have six (or was it seven?) categories of open source companies, one of which includes the Web 2.0 company that merely uses open source.

This is a way of conducting semasiology (r0ml redux): Start with an expression or word, in this case "open source," and try to figure out what we mean by it.

Now, despite the fact that I've been guilty of "mashing up" the definition of open source with Web 2.0-ish companies (like RightNow), I think I was wrong to do so. If we stretch the definition of open source business that far, then its meaning (and utility) becomes meaningless (and drops to zero). We use language to identify, among other things, and "open source business" means essentially nothing if it's used to define everything.

What we need to do, as a complement to semasiology, is engage onomasiology, wherein we start with the meaning of the word or expression and then look for the proper way to name it. (Else, Tim will end up squeezing the Web 2.0 conference with OSCON, and OSCON 2.0 and a lot less money. :-)

Open source implies more than mere use or integration of open source code into one's closed product. To me, "open source" implies liberty with the source, and not simply cost savings by free riding on open source projects. The right expression for this sort of Web 2.0 integration is "Borrowed source" or "Layer IP on top of the source code and forget about it". "Open source" really isn't a good fit, because the code isn't open (Google isn't releasing its Linux modifications, last time I checked) and isn't fundamental to how the customer acquires and interacts with the product.

Now, following the OSI open source definition, I'd like to propose a meaningful definition based on the words themselves and what we've traditionally associated with open source. At its core, the OSI definition speaks to the value of source code access and redistribution. I'd therefore define an open source company as one that has distribution, modification, and evolution of (open access) source code core to its business model and to its financial success.

Does this mean that all of its code must be open? No. Does it mean that it can't be a Web 2.0 company that heavily borrows open source code? No.

What it means is that there must be an integral, active component of contributing back source code in a publicly-accessible manner. That give-and-take of source code must be fundamental to the product and the distribution model - it cannot be an afterthought or pro bono. Or, to put it more succinctly (as Scott Dietzen says), an open source company "is a company that's business model fundamentally depends on delivering open source software."

Agreed.

However, let's also be clear that the fact that one is open source is not a magic elixir that somehow makes the business special. It just means that we've accurately defined it. Whether it's a dog of a business that deserves to implode is another matter entirely....(I only say this because I think we need to be careful to not use "open source" as a label that conveys intrinsic value. I like open source, but I don't like all open source projects or companies simply because they're open source.

Microsoft old hardware/Linux study...Market says, "We don't care"

Microsoft's Bill Hilf has just emerged from Redmond's Linux lab with this news, according to eWeek (drum roll please):

...Windows performed as well as Linux on legacy hardware when installed and run out-of-the-box, [which tests] were done in part to give Microsoft the data it needed to effectively "put to rest the myth that Linux can run on anything."

Given that I can hardly run XP (if I were to foolishly do so - just upgrading MSN Messenger on my Mac nearly killed my PowerBook) on a machine that I bought last week, much less a few years ago, I have a hard time believing Bill's lab tests. My own experience cuts against those tests.

But the larger question is, even if true, who cares? The ironic thing about all this free software is that few to none in software's largest markets actually repurpose old hardware for it. Yes, you could, but people buy new Dells to run the latest Linux - they generally don't resurrect their old machines for the task.

Bill points out in the article that this research was primarily done for the developing world, where people do repurpose old hardware. OK. I'll buy that. But Microsoft has other big reasons for losing deals (and, soon, market share) in the developing world. Having old hardware crash with Windows won't save them. :-)

Open source VC: Who's giving money?

Who are the "top VCs" when it comes to "getting" open source (and, more importantly, giving to it?). I tried compiling a list this morning (see below), and here's what I came up with. I initially just went from memory as to who has invested in what. Wanting to be slightly more scientific, however, I decided to employ some (sketchy) metrics:

To be on the list, the partner must have invested in at least two credible open source* companies.

It's actually amazing what happens when you apply this simple test. Many of the VCs that I think of as open source savvy (see bottom) have yet to achieve the volume necessary to make this list. Of course, this may be a good thing: quality over quantity. So, don't read this as a be all, end all list of who's cool and who's a loser in the open source venture world. Some of the people who didn't make the "Most active" list are great friends of mine. (But maybe that makes them a loser, even if their lack of open source investment activity doesn't? :-)



As noted, the list above is not comprehensive of all the smart open source VCs out there - just the most active. Some that I've had close interaction with, from whom I'd gladly take money (Again, this list is not comprehensive - it's early AM here in London, and I'm sure I'm forgetting people):

Ray Lane (Kleiner Perkins - SpikeSource), David Skok (Matrix - JBoss), Bob Lisbonne (Matrix - AppTran, LucidEra. Bob ran the browser business at Netscape and managed its metamorphosis into Mozilla. If AppTran were open source, he'd be on the list), Irwin Gross (Worldview - ActiveGrid), and Josh Stein (DFJ - SugarCRM).

*Disclaimer: Since no one knows anymore what "open source company" means, given that we've stretched the definition to cover any company that has ever thought the words "open" and "source" in the same paragraph, I'm applying my definition. I'll tell you what it is in a follow-on post.

CIO Mgazine: IT spending on the rise

Gary Beach and CIO Magazine are reporting [PDF] that 69.8% of CIOs are planning IT spending increases in 2006. That's great news for the tech industry, with an average 7.8% increase across the board.

Of course, stats being what they are, CIO Insight (a competitor to CIO Magazine) has its own survey of 115 senior information technology executives, and is reporting a 1.3 percent decline in IT spending in 2006 among companies with more than $500 million in sales. For those under $500 million, the survey reveals a 6% increase...and IDC and Gartner are both predicting 5% increases. Why the disparities? Baseline suggests that it may have to do with disparities in how IT is now being planned and budgeted.

Regardless of how you spin it, as with government (and seemingly everyone else), spending generally is on the rise. And never before has so much of it been allocated toward open source.

Where will they spend it? Primarily for security (50.9%), storage systems (49.7%) and lots of computer hardware (47.3%), according to CIO Magazine. (Also see this early 2005 data from Baseline Magazine.

Looks good. And, as I said, it's never looked better for open source.

Building vs. buying open source communities

Leaving aside the question as to whether one actually can buy community (see my separate post), I don't believe it's an easy task under the best of circumstances, for the same reason that most acquisitions fail to deliver their promised benefits. Combining corporate/community cultures is tremendously difficult. Combining the business models these entail is only slightly less so.

I'm an Arsenal supporter. I despise Chelsea. Part of the reason I dislike Chelsea is that it's clearly a "buy community" kind of soccer (football) club. I think the Russian mafia (ahem, Roman Abromovich) has injected nearly $500 million into the club, buying up an immensely deep bench of talented players. Oh, and I think Chelsea's coach, Jose Mourinho, is overly smug and needs to be swatted from time to time. :-)

Arsene Wenger, Arsenal's coach, has a dramatically different approach. He tends to buy potential, not experienced players. (Unfortunately, this season the result has been less than spectacular.) He has molded some of the game's great players (like Thierry Henry and Patrick Vieira), starting with them when all they showed was promise. This transfer window, instead of signing a strong player like Michael Ballack he signed two relative unknowns and one 16-year old with a lot of promise.

Arsenal builds. Chelsea buys.

Chelsea, for their part, have hardly lost a match in two years. Arsenal, after trouncing everyone two years ago, has been struggling. So Chelsea's model must be the best, right?

I don't think so. Remember the Lakers (with Shaq, Kobe, Karl Malone, Gary Payton, etc.)? The Yankees? Etc.? Each spent gobs of money on the best players, but couldn't assimilate them to make a winning team.

The open source world is no different, to my mind. In 2000, I joined embedded Linux vendor Lineo. We acquired six companies in one month (no, that's not a typo), growing from 40 to 440 in that month. We never recovered from the rapid growth (precipitated by our bankers suggesting that we had to grow fast in order to command a hefty valuation on the IPO market...right....). I learned then, and again at Novell (Ximian + SUSE), that integrating cultures and business models is supremely difficult. Some (like Jose Mourinho in soccer) are good at it. But most fail.

For new open source companies, I think the best bet is to grow a community organically. In my experience, this is what some of the smartest technology companies have done. Dan Frye and Jon Prial - both of whom collaborated on creating IBM's Linux strategy way back when - started developers working on Apache and Linux well before (two years or more) they expected to effectively monetize those projects (or, really, expected much of anything, monetary or not). SugarCRM, MySQL, Alfresco, and JBoss have shown that there is both an art and a science to commercial community building, and that it is possible and profitable.

Other companies like JasperSoft have shown that it's possible to acquire an open source project and successfully build a business around it. But I still think this is the harder task. It involves proselytizing among existing project users and developers, trying to convince them your intentions are good, rather than starting fresh. There is certainly an appeal in having a ready-made community around an acquired project, but I don't believe that keeping such a community happy is either automatic or easy.

Good technology, coupled with good marketing, will attract a community. This is true of the open and closed-source worlds. (Just look at Microsoft - they've created an incredible community of developers and users.) I just happen to think it's easier and more useful when the community is formed around an open project.

Victorian marriage and open source

I just finished Thomas Harding's Jude the Obscure, and am now plowing through his Far from the Madding Crowd. Though not necessarily central to either book, I've been intrigued by the courtship process various characters go through in them. In most instances, marriage was less a matter of love and affection, and more a matter of financial security.

For example, Crowd's central character, Bathsheba Everdene, is initially wooed by the farmer-turned-sheepherder Gabriel Oaks in this manner:

"I can make you happy," said he.... "You shall have a piano in a year or two - farmers' wives are getting to have pianos now....And [you can] have one of those little ten-pound gigs for market - and nice flowers, and birds - cocks and hens I mean, because they be useful," continued Gabriel, feeling balanced between poetry and practicality. (30)

Mr. Boldwood, the next suitor, then tries his luck on her with this argument:

"You shall have no cares - be worried by no household affairs, and live quite at ease, Miss Everdene. The dairy superintendence shall be done by a man - I can afford it well - you shall never have so much as to look out of doors at haymaking time...." (136)

Very romantic, these Victorians (especially as the men in question hadn't talked to Miss Everdene more than 5-10 minutes before making their offer :-).

I think that we sometimes treat successful open source projects this way. VCs ply project leads with promises of enterprise glory and IPO wealth. Corporations try to lure developers with the promise of temporal comfort - a comfy salary and expanded influence.

Maybe this is a good thing. I'm certainly a fan of commercial open source, as I believe that the more money developers can make from open source development, the more of such development there will be. But I wonder if the "marital offer" corporations and capital makes to the development community is sometimes a bit crass and, well, material.

I'm not convinced, anyway, that the "buy" decision is the right one in the "build versus buy" open source choice. Bringing a "free" project under the auspices of a commercial enterprise will necessarily change it. Sometimes for the better (I think, for example, the Linux kernel has benefited from enterprises investing heavily in the people involved), but sometimes for the worse (which is why it's great that open communities like Ubuntu can still persist in the somewhat commercial Linux community). For example, OpenOffice's corporate community tends to make it as bland as Microsoft Office...without the file compatibility and other essential features to truly compete with Microsoft Office. It would be better to be managed by a true community...as would Java.

Regardless of normative concerns, it's extremely difficult to weld an open source project onto the existing framework of a closed-source company. I lived though this at Novell. I think the company did remarkably well...but it still chewed up nearly two years of the company's life. Buying a community is not for the faint of heart.

No matter how much physical security you can offer that community, it may not be worth the change to that community. Something to consider.

The marathon and the sprint

At the SDForum event last week, Larry Augustin and Tim O'Reilly compared the P&L sheets of Red Hat and Borland (following r0ml's OSCON presentation), and asked, "Are they really any different?" (Very similar to the question Larry explored at OSBC San Francisco.) r0ml's answer at OSCON was that they really weren't - Borland calls it a "license" and Red Hat calls it a "subscription," but the two are functionally equivalent. The SDForum panel seemed content with this.

The problem is, they're not equivalent. One constitutes a sprint, and the other a marathon, as I pointed out.

Some believe that open source offers them a quick path to riches. Just open up the code, sit back, watch the downloads kick in, and the cash register start ringing! For any who have actually done it, however, the truth is somewhat different. It requires as much - if not more - effort to make an open source business model work than it does a proprietary one.

And a lot more patience. Open source businesses increasingly use a subscription model, which requires a monthly (or daily) recognition of revenues, contrary to the license sale, which allows upfront recognition.

So, in open source, you book revenues today but must wait to reinvest them in the business (adding employees, upgrading equipment, etc.). In the "Catholic" world of the upfront, proprietary license, on the other hand, gratification is immediate.

Hence, open source is the marathon and closed source is the sprint.

A sprint is nice if you want to get the run over with quickly. Sell to one customer, and move on to the next. If you're the customer, of course, this model stinks, but who cares about customers?

Open source companies have to. Their model requires it. At Alfresco I get paid a little each day by Informa, Boise Cascade, and our other customers. I have to earn that by providing super attentive service, a superior and evolving product, etc. At no point can a I (or Sugar, or MySQL, or Red Hat, etc.) rest on my laurels, celebrate the sale, and forget my customer. (Someone at the SDForum event suggested that Red Hat only has a 50% renewal rate, which I had heard before. If this is true, I think it doesn't necessarily speak ill of the model but rather points to an area that Red Hat can continue to improve upon. Their explosive revenue growth suggests that they're doing just fine in that department.)

So, it's a marathon. Not always pleasant (as I learned running my first marathon a few weeks back - my right knee still won't forgive me), but better for customers and better for vendors, too, once they're on track and on pace. It forces us to think about the long haul. And, not to take the analogy too far, but whereas in a sprint you can dope up (take shortcuts) to boost performance, there really is no way to improve one's marathon ability...except by putting in a lot of miles. There's no shortcut in a 26.2-mile race.

Microsoft's internal open source community

More often than I'd like, I'm forced to acknowledge "mea culpa" and reverse course on something I've said. I've always been more optimistic than most on Microsoft's chances to get open source right, but I occasionally think I've met exceptions to the rule.

I thought I had discovered one in Sam Ramji. He's part of Microsoft's Emerging Business Team, and I was fortunate to share a table with him at the recent SDForum event. Like virtually everyone I've met at Microsoft, Sam was articulate (sophistic is the word I sometimes apply to Microsoft friends like Jason :-), intelligent, and not afraid to back down from an argument.

We had plenty at our breakout session. (I learned after the fact that I wasn't supposed to blog this session. Again, mea culpa. Since Google doesn't forget anything, there's no point in trying to erase what I wrote. But I can at least acknowledge when I was wrong.) In the case of Sam, I misunderstood both what he said and what he meant.

The discussion went round the table about whether open source is a follower or a leader. I've taken both positions in the past, but generally believe that commercial open source obviates community open source's (not really a true distinction there, but I assume you know what I mean) definitional need to follow.

(Open source "in the wild" relies on a sufficient body of developers with interest and aptitude for a given technology, which happens once the groundwork has been laid by someone else. So, for example, you would not likely have an open source database until databases were well understood and a large enough group of developers could combine around creating one. I suppose one developer could innovate something on their own, but the money required to build up a mass of skilled developers around a technology largely precludes it, in my opinion.))

Anyway, I took Sam's comments to be pejorative on this, but after exchanging emails with him, I think I was wrong. Sam was no more antagonistic toward open source than I am. In fact, his background (lots of open source in it) makes him as likely as I am to be friendly about open source, and understanding of it.

Sam's response to my misunderstanding? He sought me out and tried to understand why I had misunderstood him. This has universally been the response from Microsofties that I've known. It's one reason that I put Microsoft in a very different class than most companies threatened by open source. The company started out as belligerent toward open source (and, indeed, sometimes this old tendency flares up), but quickly learned that this was the wrong way to become part of the community.

I'm actually quite bullish that Microsoft will continue to figure it out. Heresy, I know, but I'm not religious about software. I haven't met anyone there yet that hasn't shown enough humility to truly become part of the community (then again, I've yet to meet Ballmer or Gates ;-). If Tim thinks Microsoft could well be defined as an "open source company" because of their code contributions, I can agree, too. But I think the greater test is the quality of the people in the company. In this department, Microsoft has a very deep bench.

Sorry for misunderstanding you, Sam. Thank you for reaching out to me. And thanks for being a credit to your employer. Sorry I said otherwise. I was wrong.

And now I will go back to killing your SharePoint business. :-)

Tim O'Reilly: What business is Red Hat in?

I'm sitting in SDForum's Open Source in the Enterprise executive summit. Great so far, with a strong turnout.

Tim O'Reilly just asked a provocative question, and has repeated it several times. It is:

What business does Red Hat think it's in?

Tim's rejoinder to his own question is, "They're an integrator, not a technology company. They should be Dell, not Microsoft."

He has a point. Red Hat has become what it is by integrating and delivering the various components of the Linux operating system. An intriguing and (mostly) valid argument. It doesn't take into account Red Hat's new R&D push, but it does capture what Red Hat does well: deliver an excellent community-driven product.

Microsoft and its open source partner ecosystem

Today, it's just JBoss that is a formal partner with Microsoft. (And maybe MySQL? I don't know....) But tomorrow, many more?

So says Bill Hilf, Heir of Matusow (no, it's not quite the same as being the Heir of Slytherin ;-) in this eWeek article:

Expect to see a lot more interoperability work between Microsoft Corp. and some of its open-source competitors over the next year—like the agreement struck with JBoss Inc. last year—as well as more participation by the Redmond, Wash., software maker in preventing interoperability problems earlier in its product cycle and providing potential fixes when issues arise.

"We have been successful in identifying popular open-source software applications that our customers are interested in using on the Microsoft Windows Server platform and working with those companies or projects to ensure that solution is well integrated," said Bill Hilf, the director of Platform Technology Strategy at Microsoft, and also the man who heads the company's Linux and open-source lab.

I'm intrigued. But given that I represent the death of Microsoft's SharePoint business, perhaps they won't roll out the red carpet to Alfresco? :-)

Big company = Big bureaucracy = Big opportunity (for open source)

I've written before on the value open source brings in lowering the bar for trying out software. Today, I felt it painfully.

I had conversations with two (very) large companies, both of which wanted to do a trial of Alfresco's Enterprise product. Simple, right?

Well, no. It seems that the companies require a trial agreement to be signed. With one, they only sign on Mondays, so they have to wait a week to try it out (unless one of them wants to just download it and try it out - I won't tell :-). With the other, I was told it could be months before their Legal department would sign off on it - not because it's open source, but just because that's how they operate. I begged him to just download it and give it a spin, but he couldn't. So they're going to try out the hosted demo, but still....

Months?!?

We desperately need to rethink how we access, trial, and buy software. The old way just doesn't work.

Open Source Work Ethic and the Spirit of Capitalism (Red Hat et al.)

My wife and I live in a "cute" neighborhood. Unfortunately, "cute" means "old Tudor-style houses [ours was built in 1941] that tend to be small and expensive." With four kids and me working from home, she was at the end of her tether and declared, "We've got to get into a bigger house!"

My response? "Well, that's very Catholic of you."

Now, lest you think I was engaging in religious baiting, I was merely making a reference back to Max Weber's seminal book, The Protestant Work Ethic and the Spirit of Capitalism. (That and I was just trying to antagonize her. :-) If you haven't read it, you should.

In it, Weber asserts that the Puritan/Protestant religion tended to fuel capitalism, and that Catholic beliefs tended to dampen the spirit of capitalism (though not the desire to make money, itself). From Wikipedia:

Weber shows that certain types of Protestantism favoured rational pursuit of economic gain and that worldly activities had been given positive spiritual and moral meaning. It was not the goal of those religious ideas, but rather a byproduct - the inherent logic of those doctrines and the advice based upon them both directly and indirectly encouraged planning and self-denial in the pursuit of economic gain.

Weber traced the origins of the Protestant ethic to the Reformation. In his opinion, under the Roman Catholic Church an individual could be assured of salvation by belief in the church's sacraments and the authority of its hierarchy. However, the Reformation had effectively removed such assurances for the average person....

In the absence of such assurances from religious authority, Weber argued that Protestants began to look for other "signs" that they were saved. Worldly success become one measure of salvation...[and]...Weber saw the fulfillment of the Protestant ethic...in Calvinistic forms of Christianity.

The "paradox" Weber found was, in simple terms:

• According to the new Protestant religions, an individual was religiously compelled to follow a secular vocation with as much zeal as possible. A person living according to this worldview was more likely to accumulate money.
  
  
• However, according to the new religions (in particular, Calvinism), it was considered a sin to actually spend this money on personal luxuries or on religious icons. Also, charity was generally frowned on because a lack of worldly success was seen as a combination of laziness or divine disfavor.

The manner in which this paradox was resolved, Weber argued, was the investment of this money, which gave an extreme boost to nascent capitalism.

Whether valid or not, a truly interesting take on history and economics.

How does this apply to open source? I remember having a conversation with several members of Red Hat's executive team about its business model, which thoughts were later echoed in a conversation I had with a Red Hat board member. Red Hat's business model very much is a "save now, recognize revenue and spend investment later" type of model.

And it is diametrically opposed to the traditional software business model: recognize huge license fees upfront and spend them with abandon. Red Hat (as well as other subscription-based businesses like Alfresco, MySQL, JBoss, and SugarCRM) may be able to book the full value of a deal upfront, but it actually recognizes the revenue on a daily basis thereafter (others, like Alfresco, still recognize revenue on a monthly basis). The value of the model is that, once it kicks in, it's extraordinarily powerful - a recurring, predictable revenue stream upon which the company can depend and invest against. But there's a lot of "Calvinistic" saving upfront to get the company to that point.

Furthermore (still beating the Red Hat drum), Red Hat does a fantastic job of using its Fedora project to "read the tea leaves" of what enterprises will want, and then delivering those in a stable, enterprise (RHEL) version down the road. As a result, the company was last to the 2.6 kernel game, among other things. It takes its time, makes its investment, and then releases the product. The result? It's one of the top-ranked companies in all of technology, and has a 97% approval rating with its customers. Save and invest now, spend later.

If you want to succeed in open source (or, really, in any business), this must be your mantra. There is no supernatural authority that will "save" your business. There is only the drudgery of working your tail off today, and banking that investment today, so that you can spend it on further drudgery tomorrow. Not very cheery, but the payoff is huge.

In closing, it's worth hearing Weber directly on this work ethic:

...[T]he summum bonum of this ethic, the earning of more and more money, combined with the strict avoidance of all spontaneous enjoyment of life, is above all completely devoid of any eudaemonistic, not to say hedonistic, admixture. It is thought of so purely as an end in itself, that from the point of view of the happiness of, or utility to, the single individual, it appears entirely transcendental and absolutely irrational. Man is dominated by the making of money, by acquisition as the ultimate purpose of his life. Economic acquisition is no longer subordinated to man as the means for the satisfaction of his material needs. This reversal of what we should call the natural relationship, so irrational from a naive point of view, is evidently as definitely a leading principle of capitalism as it is foreign to all peoples not under capitalistic influence. At the same time it expresses a type of feeling which is closely connected with certain religious ideas. If we thus ask, why should "money be made out of men", Benjamin Franklin himself, although he was a colourless deist, answers in his autobiography with a quotation from the Bible, which his strict Calvinistic father drummed into him again and again in his youth: "Seest thou a man diligent in his business? He shall stand before kings" (Prov. xxii. 29). The earning of money within the modern economic order is, so long as it is done legally, the result and the expression of virtue and proficiency in a calling; and this virtue and proficiency are, as it is now not difficult to see, the real Alpha and Omega of Franklin's ethic, as expressed in the passages we have quoted, as well as in all his works without exception.

Matt repents in sackcloth and ashes (Skype)

OK, so I haven't been the biggest Skype proponent thus far. I never believed that businesses would use it as a serious tool, or that average consumers would use it for communications.

Today, however, my IPEVO free.1 Skype phone came, and I've been having a great time trying it out with my Skype buddy list. (Many thanks to Om for pointing out this first Mac-friendly Skype phone. For those still relegated to Windows, there are several more options, which you can find on the Skype website.)

It's completely integrated into Skype, such that I can scroll through my buddy list, dial a call, etc. from my phone. The downside is that it's a USB phone, not wireless/cordless, but I'm sure this is something that will be overcome in the very near future.

In the meantime, I'm finally able to comfortably take an incoming Skype call, rather than scurrying for a headset. I'm looking forward to it.

(Note: I still think Skype and others that rely on someone else's network are going to get a wake-up call at some point. But for now, it's nice that its free. By the time it's not, it may well be worth paying for.)

Self-congratulatory news: Alfresco's PHP support

I normally don't do this, but since I'm a little tired of hearing "Java is dead!" from Dave, Peter, and crew :-), I thought I should note that Alfresco just announced support for web scripting languages like PHP. We've actually always had this, but never advertised the fact. Basically, it means that developers don't have to learn Java to work with Alfresco. They can extend the system with the web scripting languages they know and love.

This, btw, is how all systems should be architected: let the core development team state its preference (in our case, Java) and write it as such. But also ensure that third parties who don't share this preference can modify and extend the system in the language of their choice.

Anyway, here's the press release (also see below for excerpts), and here's where you can download the PHP module.

Alfresco Announces PHP Development Interface

Jan 17, 2006

Leading Open Source ECM Platform and Leading Programming Language Integrated

LONDON-January 17, 2006-Alfresco Software Inc., the first provider of an open source enterprise content management solution, today announced that has integrated the open source programming language, PHP, into its development environment. This enables PHP developers to create new content-centric applications and dynamic web pages that access Alfresco....

"PHP has a long heritage in open source," said John Newton, chief technology officer, Alfresco Software, Inc. "“We wanted to provide an open platform for Enterprise Content Management for all languages and development environments. With the architecture we have adopted users can get the best of both worlds - the ease of development with PHP and the enterprise robustness of Java for open source ECM."

"The promise of Web Services is that developers can create sophisticated applications without having to recreate the underlying complex technologies such as content management. Alfresco Software is one of a growing number of companies that recognize the power of PHP to expose the power of their technology via Web Services," said Doron Gerstel, CEO at Zend Technologies, the originators of the PHP language. "Providing a PHP-API will enable their community to create tailored content management applications for vertical industries such as publishing, healthcare, or engineering. Zend is committed to collaborating with Alfresco and others to make PHP the most productive platform for creating applications that integrate such Web Services."

So, you can have your cake and eat it, too. That's the benefit of a clean, extensible architecture: you can be popular with the O'Reilly web scripting crowd as well as the Jonathan Schwartz Java crowd. :-)

Red Hat: The mother of all open source business models

The biggest problem in open source today is that few know how to monetize it. There are great developers out there who will write free (as in freedom and price) software regardless of a profit motive, but they are the minority. These generous benefactors are not sufficient to fuel the continued rise of open source software.

So, we need more developers making more money to fuel...more development. But how? There are a few promising models (and the Open Source Business Conference is the best place to learn and develop new ones), and many more that are complete rubbish.

I'd like to focus on one, in particular, that I think is absolute genius: Red Hat's. Oddly enough, I can't think of a single other company that uses Red Hat's model, despite its evident success. I'm not sure why, except that I do think few understand the nuances of Red Hat's model. I'd therefore like to spend a few minutes trying to unpack exactly how Red Hat operates.

Red Hat's model is a product of necessity. The company had to figure out how to survive its lack of code ownership, and found a brilliant way to turn this apparent deficiency into a strength. In turn, this has allowed Red Hat to invest in innovation that all benefit from, but which benefits Red Hat particularly. (As a related side note, without a good business model, open source and innovation are at odds - a good business model will allow a company to heavily innovate, secure that it will be able to profit from such innovations as much as and more than its competitors.)

It's a bit like Arsenal allowing Thierry Henry (shown here pummeling Middlesborough in January 2006, 7-0) to play for opposing teams, knowing that the goals in Arsenal's favor will always outweigh those against it.


Here's how Red Hat's model works:

1. Red Hat helps to drive Linux development, ensuring momentum and competitiveness for the open source project.
   
   
2. Red Hat splits its Linux offerings into two camps: Enterprise (Red Hat Enterprise Linux - the one you'll see prominently displayed on the company's website) and Community (Fedora). Attention and energy is primarily focused on the product that will actually bring in revenues (RHEL), but Red Hat is careful to also nurture the development community, which will pay it little to nothing but which brings other, less tangible benefits.
   
   
3. Red Hat tests and certifies RHEL to run on certain hardware, with certain software. Red Hat restricts access to this certified, supported RHEL - you can get the raw source for the uncertified RHEL, but not the compiled, ready-to-go binaries. Only paying customers get that. (Note: Red Hat recognizes that few to no large enterprises are going to depend on an unsupported, self-compiled distribution.)
   
   
4. Red Hat ties support to its software - you cannot get and run the RHEL binary noted above without buying commensurate units of support. Red Hat ensures this through its ingenious subscription agreement. If you want the real Red Hat, you must pay - there's no effective way around it. (There are workarounds, but they're not worth the bother. Red Hat knows this, and mints money from the result.)
   
   
5. Red Hat delivers updates (and ensures customers stay with it) through the Red Hat Network. Companies plug in, get updates, occasionally call for support, and make Red Hat an explosive, important open source company.

That's it, in a nutshell. The important things to remember are:

• Red Hat's model works because Linux is successful.
  
  
• Red Hat ensures Linux's success by investing heavily in it, and marketing it heavily.
  
  
• Red Hat does nothing to prevent would-be customers and developers from trying it out (quite the opposite).
  
  
• Red Hat makes it hard to impossible to get the compiled, binary version of its tested/supported/enterprise-ready software without paying it. (A recognition that while source is free, few actually want source, and even fewer pay for it.)
  
  
• Red Hat ensures an ongoing relationship (and payment channel) with its customers through a subscription agreement.

Sheer brilliance. Red Hat is able to deride competitors who have hybrid models while effectively having one itself, scoring marketing points and positioning itself as the one, true open source Linux vendor. Red Hat is able to lower the barriers to trying out its products without abandoning the ability to convince customers to pay upfront and ongoing support/license fees. Red Hat is able to make a truckload of money.

The only question at this point is, why aren't you using the Red Hat model, too?

John Mark in the garden of good and evil

John Mark has at least two things going for him. He likes soccer - a real sport - and he writes for a long time. (Now he just needs to start following real clubs...which means he'll have to look beyond the MLS. I've got a good team to share with you, John Mark....) His first salvo post-new LinuxWorld gig is mighty long. So long even I started skimming, and I read a lot of novels. :-)

John Mark makes some good points. (Evidently so, as even the august Nick Carr picks up JMW's tome and runs with it.) He makes them forcibly, though sometimes at the expense of accuracy. That's OK - I like good soundbites as much as the next guy, and I think John Mark has both substance and style. (Ever listen to Larry Lessig? Lots of substance, but sometimes he employs style to bridge some logical gaps, as we used to chide him for in class.)

The core of John Mark's argument is as simple as it is subject to debate: open source just is, and would have happened no matter Linus or RMS because it's just a symptom of the natural order of things: commodification. The Internet only exacerbates (or facilitates, if you will) this trend.

Various people (including me) have made this argument for several years. Open source is a weapon, a tool. It is an efficient distribution methodology. It is a way to undercut competitors (as IBM has done masterfully) and a way to enable customers to share in the development process.

But it's not necessarily the most innovation-inducing way to write software, as John Mark argues. He says:

This continuously evolving collective knowledge base carries another consequence: speed of innovation. Because of the speed with which users, developers, and companies can post documentation, patches, or new software projects, the product life cycle has shortened considerably. Software vendors must work harder than ever to stay ahead of the floating software boats. This constant drive for innovation means that products released just yesterday lose value more quickly than before, due to future products already filling the software pipeline.

I don't think it's true that the pace of innovation has increased - instead, I think the visibility into that innovation has improved.

And I certainly don't believe that product lifecycles have shortened. Those (like Red Hat) who tried to behave otherwise were told clearly by the market that it likes a nice and easy pace, thank you very much. Microsoft and the proprietary vendors could release as early and often as the open source vendors could, but recognized long ago that customers don't actually want this.

In some areas, like the 'Net, Google and other closed-source vendors have found that a perpetual beta mentality can work, illustrating that it isn't open source, per se, that speeds innovation, but rather that the Internet enables a fast, incremental product release schedule...if desirable. Generally, it is not. So, again, it's not that innovation has increased its pace, but rather that we have greater visibility into what innovation is happening, whether closed or open.

I also don't believe that open source is necessarily The Best Way to write software, contrary to what John Mark says. (I think it's far too soon to associate once-and-for-all superlatives to open source. Better? Sure. Best? I'm not yet ready to stop the evolution of software.) It simply may be the best way (so far) to ensure written software is fixed. (You see the Jamesian in me coming through - pragmatism.)

John Mark says:

With prices approaching zero, software developers have two choices when trying to win over users: (1) add features not available elsewhere, and (2) release the source code. There is no other currency of value that developers can extend to users. In fact, these two options are actually interrelated. A software developer trying to accomplish option 1 on his own will face a daunting task, whereas a developer who releases source code, assuming the project is viable, will have a ready supply of suggestions for improving the software and adding features. She will also have a cheap and fast means of distributing the software complexity around the world, with the implied deepening of the collective software knowledge base. This collective software knowledge base will pay dividends in the form of knowledgeable users and contributors.

Well, maybe. I'd feel more comfortable with his opinion on this had he kept at it with GroundWork and proved the financial superiority of the model. :-)

I'm also not sure that John Mark's assumptions closely correlate to the reality of how open source projects work. Given some other comments John Mark makes in his article, I know he recognizes this. But he conveniently overlooks this in an effort to make a point stick.

This actually ties into an overly strong, and not quite accurate, statement that John Mark makes: "there is no open source community." This isn't true. There is an open source community. Lots of them, actually. Alfresco has one. So does SugarCRM. Etc. Open source is a community of small, sometimes niche-y communities. John Mark is right that there isn't any one open source community, but it's not true that there isn't one at all.

Again, John Mark knows this. He actually alludes to it in his argument, holding that those who build such communities around their projects will have competitive barriers to entry, as I've argued as well. Having said that, and I'm sure John Mark would agree, this kind of project/product user community phenomenon really isn't any different from what successful companies (like Microsoft) have always done. Maybe what John Mark should have said, then, is that open source is not a special kind of community, and it's not a monolithic community. I'd buy that.

Well, maybe I wouldn't. That's another problem I have with John Mark's argument, but one he promises to answer in a separate article. That is, he talks about how open source is a superior economic model, but he points to no concrete evidence that this is true. I had lunch with two fellow open source entrepreneurs today, and it was clear on both sides that there are lots of things that open source makes easier...and many (like making gobs of money) that it makes somewhat harder. Or, at least, different. I'm looking forward to his second article.

Despite these quibbles I have with John Mark's article, I'm glad he wrote it. We need more logic and reason in open source, and less dogma and blind faith. I assume it will be John Mark who takes over OSBC later this year - it's comforting to see the conference will be in good hands.

Seeing is believing: Scalix

Julie Farris did me the honor of demo'ing the Scalix web client this morning. I've known about Scalix for years but, incredibly, really didn't know exactly what the company did. I knew it was a Linux-based email server, but didn't know that the company's web client meets and sometimes exceeds the sexy UI/functionality of Zimbra, Oddpost, etc.

Unlike other offerings, I really couldn't tell that I was in a browser, because they yank out the browser artifacts (most of them, anyway) so that there is no back button, address bar, etc. It's a very simple tweak, but important for making fat-client lovers like me feel comfortable. I also really liked the windowing system (very similar to CanyonBridge's offering, which I think is fantastic).

I guess this is ample reason to attend OSBC (you knew a self-congratulatory statement was coming, didn't you? :-). It's nice to talk about what open source can do, but OSBC is becoming the DEMO of open source applications and middleware. At this one, we'll be showing off 12 leading open source products/projects with the entire crowd (between keynotes) looking on. I'm hopeful that through public demos like this, my ignorance (and others') can be dispelled.

Up until I actually saw Julie demo Scalix, I didn't think much about the company, because (empiricist that I am) I couldn't see/feel/touch the software. Now that I've seen what it can do, I'm pushing Alfresco to actively consider it as an option for our email upgrade. I'm guessing that many of us are like this with open source - we like it in theory, but often don't have concrete examples of killer software to point to. I'm hoping that OSBC will become a forum for demonstrating that open source truly has arrived and is in some cases clearly superior to proprietary software.

Thanks, Julie!

The 451 gets something wrong: the nature of community

As conventional open source wisdom goes, the below citation from The 451 Group's December 2005 report, Cashing in on Open Source is about as perfect as you can get. Every open source developer will pat herself on the back as she reads it:

Both pure proprietary and feature-constrained models [i.e., company that has open and closed versions of a product which are not compatible] embody the unwarrantable assumption that your internal, paid developer community will consistently do a better job, over time, than their counterparts in the open source community. It's possible that they can, but on balance of the evidence presented to date, we doubt that they will.

The problem I have with it is that it belies a belief in The Open Source Community, which is a bit like a belief in Santa Claus. There's some truth to both beliefs, but also a lot of untruths.

For one thing, there is no such thing as THE open source community. It's a community of communities. For another, most open source projects are either non-projects, non-functional (the vast majority of SourceForge "projects" are nothing but moribund signals of intent, as has been shown numerous times), or non-community (most SourceForge projects have just one developer, and the huge majority have fewer than five).

Right now, the biggest issue in software is how to attract a user/development community, and not how to appeal to THE open source community. That's why we're heavily focusing this February's Open Source Business Conference on the subject of community-building. And it's why The 451 Group's comment above is a bit too glib for my taste (and not in keeping with the normally concise and meaty analysis it normally dispenses).

Open source projects - every single one of them - rely on a small group of developers, with a wider body of bug fixers (but still quite small), and a comparatively large body of bug finders. This largest body is the user community. Successful open source companies and projects are those that find ways to grow this community, and not those who (vainly) try to attract significant external core development. Such development happens, but not very often. About as often as Santa Claus visits.

The 451 Group: No $ salvation in going open

There is a myth out there that open source offers a convenient way to save a dying company's proprietary product, or provides a way to head off the eventual bleeding dry of a successful closed-source product (like Oracle's database, for example). The 451 Group has news for you:

Nope. Sorry. Not going to happen.

In its Cashing in on Open Source report, they make the following point:

Pure plays commodifying new sectors and offering professional support may be profitable - some are remarkably so. Examples include the Linux distributors, JBoss and MySQL. Traditional companies, however, should not expect open source initiatives to contribute substantially to the bottom line. The cost of building a developer community (Eclipse, OpenSolaris, openadaptor, eBay) is a sunk cost. Returns are intangible, such as grassroots marketing or a thriving ecosystem of third-party software. For traditional software companies, open source threatens to erode existing revenues, but it cannot promise to replace them.

While I think a more judicious use of the word "remarkably" is in order, I otherwise agree with the sentiments expressed here. I know from my time at Novell that adding open source to a dwindling NetWare business didn't magically transform it into a vibrant, growing NetWare/OES (Open Enterprise System)/Linux business (though I continue to believe the company is on the right track here). Community doesn't automatically equal sales, in other words, and certainly not the kind of sales traditional software vendors are used to squeezing out of their closed binaries.

Which brings up an interesting point. I think it is structurally, systematically impossible for a closed-source vendor to replicate these closed-source fortunes in open source. It simply will not happen. Open source requires a different sales and marketing mix, lower overhead, etc. In short, a very different business model. No closed-source company is going to be able to make such dramatic changes without being pushed to, and at the point of extinction (which is basically the only thing that will drive these companies to give up the license umbilical cord), who cares what the company does anymore?

There is hope, however. But it's not in replacing one's closed source business with open source code. Rather, it's in augmenting a dwindling closed source business with a rising open source business, and gradually changing out the old organization that supports the old, proprietary business with a new, open source friendly (business model, culture, etc.) business and associated product line. Look for adjacent markets and start building out an effective open source presence there, rather than trying to pretend that you're truly willing to cannibalize your remaining profits in an existing business. Few are. Maybe none.

The 451: The value of source

The 451 Group's Cashing in on Open Source report contains an off-hand comment that I think deserves more comment. It says:

Because human-readable programming languages compile to machine-readable binaries, it'’s possible to sell usable software without the source code attached.

What The 451 doesn't point out is that the inverse corollary is also true: because most people don't care about source code, and because compiling from source to a binary executable is work/drudgery, you can make a business out of distributing free and open source code and charging customers a fat fee for a "certified" binary.

Don't believe me? Take Red Hat as an example. It makes its money by charging customers to get the magical, certified Red Hat. In truth, it's the same RHEL you can download for free, but the assurance of certification (safety!) and forced support (you buy it whether you want it or not on a given server - see its licensing terms here). I think Red Hat is one of the most sophisticated software companies out there, and this license and business model is proof. Red Hat's model is not JBoss', MySQL's, or anyone else's. In fact, I know of no other company that replicates its model. Why? The proof is palpable....

The open source people network (451 Group analysis)

I'm still sifting through The 451 Group's findings (from its Cashing in on Open Source Software report), and came across something (page 94, for those of you following along at home) that seems like an absolute gem of an insight. The 451 Group writes:

There are plenty of nonmonetary benefits for corporate source-code donors as well. Programmers can preserve their investment in their own work when they move from company to company. This sounds unattractive on the surface, but it has shown benefits in some industries - investment banking for one - where a small number of highly qualified people tend to flow back and forth. Some programmers will take knowledge with them when they leave, but the value lost may be more than compensated for by the fact that releasing internal code 'into the wild' lets external developers train themselves in the software and become potential expert hires.

One way to look at this is a way to make one's in-house developers fungible. If Developer X leaves, no big deal, because her code is open source and someone else can quickly come in to learn it and then extend it. Since non-competes don't work in the most important tech center on earth (California), it's a way for companies to preserve their investments...through disclosure, rather than through stealth.

Another way to look at the issue is from the developer level. This is a way for a developer to maximize her value to the wider technology network. Their talent becomes fungible in a way that is at best crassly approximated by the closed source hiring network. If I'm a rock star developer for Google, I have to be really big name internally to have any resonance outside the company. But if I'm an open source developer, my talent is broadcast to every relevant tech company (or customer) out there.

If I were a developer, I know which I'd prefer. The latter course would mean I'd have a steady stream of offers. The former course would require me to actively sell myself, network, etc. No contest. Open source wins.

Either at the employer level or the employee level, open source should be a more efficient and effective way to maintain an investment in talent and technology: one's employees or oneself.

My 15 seconds of OpenOffice (in)fam(e/y)

It always surprises me how knee-jerk we can be in the open source world. If someone writes a glowing report of Project X, we slobber all over ourselves to salute them, however inaccurate they might be. If someone indicates a need for improvement, we slam them.

My recent post on OpenOffice is a case in point. I just don't think it's very good, and I also don't think it matters that it's not very good (because my goal in life is not to beat Microsoft at its game, but rather to make Microsoft play on my turf. Here are a few reasons OpenOffice is weak, and is not likely to ever attain Apache/Linux/BIND/etc. status.

• OpenOffice isn't a piece of software that most developers care about, and developer-aligned open source software is much more likely to be superior software. If there's no itch to be scratched with OpenOffice (and most developers don't spend their time fretting about presentations, spreadsheets, and such), then it's not likely to be heavily reviewed and developed.
  
  
• In part because of the above-mentioned point, OpenOffice has a small community. Sun and Novell. There's very little outside development, and the project isn't the top priority for either company. Yes, commercial open source projects can be highly successful (MySQL, Alfresco, SugarCRM, etc. are good examples), but they depend upon companies with a financial interest in making them work. Sun doesn't go out of business if OpenOffice doesn't take off. Neither does Novell. But SugarCRM does if its product and associated community fails. That's the difference.
  
  
• OpenOffice is very hard to develop, unless you're an active, deeply involved contributor. It's just too big to invite "drive-by development," and isn't modular enough to allow casual development from passersby. It started proprietary and monolithic. The proprietary part has been solved; the monolithic part? Still there.
  
  
• At the end of the day, who cares if OpenOffice is successful? I think some of the vitriol poured out in the commentary here and elsewhere has to do with an inferiority complex of sorts, which I think is bizarre. Why does open source have to do everything equally well? And who cares about an office suite, anyway? I spend 95% of my day in email or a browser. In these areas, open source is actually superior (FireFox) or nearly so (various email programs). I don't care very much about my office suite. It's something I use when I have to, but I keep it to a minimum, because it's not built for distribution and connection, whereas these other tools are.

So, in short, let's please spend our time doing something other than fretting about what I or anyone else says positive or negative. If it works, use it. If it doesn't, don't. OpenOffice isn't manna from heaven; Linux probably is. One can be paltry with the other very strong without open source, as a development and distribution mechanism, being injured in any way.

The problem(s) with OpenOffice

Before you start flaming, please understand that both I and the author of the news article are open source PROponents. We like open source. I've made a career of it.

But that doesn't mean open source is somehow infallible. Sometimes the comments here read like a devout Catholic thinking I've started a smear campaign against the immaculate conception. Keep in mind that it's just software and that it can always get better. Things get better as we find their bugs and improve them.

So, with that caveat....

OpenOffice stinks. Though it has become much better since I started using it back in 2000, it's still nowhere near what it needs to be to effectively displace Microsoft Office. (I've questioned whether we should even care about displacing it, but that's a separate discussion, and one that largely makes this whole dumb ODF debate pointless.) Why?

A few reasons. As Andrew Brown in The Guardian notes:

The myth of open source rests on two improbable assumptions. The first is that a significant proportion of users can fix bugs. That is true at the Massachusetts Institute of Technology, where the concept of open source was first formalised in the 1980s by Richard Stallman and others, and it is true in some of the geekier corners of the internet. But on programs intended for use by the non-programming public, it's a very different story.

This is important because of the second crucial false assumption: that even if not all users can fix a bug, they can help find them. They can't. Most users just think: "The computer isn't doing what I want."

I've talked about this before, too: open source doesn't effectively pull in the collective intelligence of the code Proletariat. You won't find my parents, or even me (allegedly tech savvy) contributing to open source projects. We're technically incapable of doing so. And yet, we are the primary users of things like OpenOffice. Where's our vote?

Even those who are capable of fixing bugs and adding features to OpenOffice don't necessarily do so, and many can't. Why? Either they don't have time or inclination (their "itch" is elsewhere), or they don't have the ability to do drive-by development on a gargantuan pile of code like OpenOffice. Says Brown:

But what about the innumerable volunteers who can download the code and fix what they like? They take one look at the effort involved and run. OpenOffice is an extremely complex mountain of source code. As far as I know, in the five years it has been available as open source, not one contribution to the program has come from amateurs. The outsiders who have provided input have been full-time professionals employed by Linux companies to help make the software credible.

There has been a lot of volunteer effort, but it has gone into support....[B]ut the overwhelming energy [around OpenOffice development] seems to go into filling the blogosphere with remarks about the merits of open source software and getting outraged about inconvenient facts.

What facts? Well...(again, Brown):

So why is OpenOffice so dire? The project claims more than 50m downloads of the software, so let's assume that 50m people have tried it at least once.

More than 50,000 bugs have been reported. And how many have been fixed by open source's uniquely efficient processes? According to the (public) bugs database, at last count, there were more than 6,000 unfixed bugs, and more than 5,000 feature requests. While the number of bugs discovered seems to rise with the number of users, the number of fixes doesn't, and the number of fixers certainly doesn't. Only about 500 people have signed the legalese that would enable them to submit code to the project; since you need to do this even to make changes to the website, that will translate to far fewer than 500 volunteers submitting real code. A reasonable guess would be 50, or even five.

Meanwhile, there are some simple, hugely irritating bugs that are four years old. Two obvious ones: notes (or comments, as Word users call them) don't have word wrap; and spaces typed at the end of a line won't show. It's not many eyes making bugs shallow; more like many eyes making bugs invisible.

Most software has similar irritations. But complex open source projects seem uniquely badly placed to fix them. They rely on a very small group of programmers relative to the user base, and who have no direct incentive to work on the bugs that are important to users.

I don't necessarily agree with Brown on this last point. His comments may be true of "pure" open source development communities, but they are not accurately reflective of commercial open source projects, which benefit from a "pure" community effect but also have a corporation behind the code to take care of the ugly little nits in the code that free-love developers won't.

So, is the answer to throw out OpenOffice? No, of course not. Rather, we need to acknowledge its defects and work to fix them. How? Well, I'd personally recommend that Sun or Novell take on the project as a serious commercial offering. Sun comes closest to this with StarOffice, but I think the company would feel more inclination to improve it if it derived unique benefits from doing so. Arguably, it does today, but I think we could amplify those.

Would this make it less of a "community" project? Not really. Sun already does the vast majority of OpenOffice development. Such a move would simply recognize this fact and leverage it for Sun's benefit...and ours.

Back to the future: Blogging The 451 Group's open source report

The other day I noticed a frightening thing: I was starting to dress like my six-year old son. Isaac is a great kid, but he tends to make weird clothes combinations, combinations that I subconsciously was starting to find cool....

In similar manner, the software industry's BigCo's are increasingly dressing up like the startups bent on supplanting them. I've noticed this before, but it wasn't until I started reading The 451 Group's excellent report - Cashing In on Open Source - that it really hit home. In its opening paragraphs the report authors state:

...[P]erhaps the most remarkable finding of the report is the extent to which ideas and methods promoted by open source evangelists five years ago are now treated as conventional wisdom, especially by such pillars of the 'open systems' Unix and J2EE culture as IBM and Sun, and even by Microsoft itself. This explicit acknowledgement of the merits of source-available development by the epitome of closed source software vendors, more than any other single factor, persuades us that open source software is here to stay.

This is a fascinating insight, and indicative of the kind of fresh thinking that The 451 brings to the industry. No tired, cliche'd thinking here, and plenty of at-times irreverent, always incisive thinking. In short, not much like most analyst reports out there. :-)

Over the next week, Dave Rosenberg and I are going to be blogging insights from The 451 Group's report. I'd highly recommend that you purchase it.

LinuxWorld goes OO.o

Here at Alfresco, we received word back from John Mark Walker, the new program chair for LinuxWorld, that our session has been approved and we'll be speaking at LinuxWorld Boston. As usual, we have to sign a speakers agreement. (These, btw, are the bane of my existence. IDG always makes us distribute these to OSBC speakers, and I find them as annoying to collect as they are pointless to enforce. But you didn't hear me say that. :-)

But not as usual, John Mark sent the agreement in OpenOffice format.

This struck me as odd, given that PDF would have been an easier format for 99.999% of the world to handle. Or .doc, for that matter. Fortunately, I have NeoOffice/OpenOffice installed on my Mac, so I'm blessed and can open the document. But I'm betting a fair number of LinuxWorld speakers will be flummoxed by the format and waste an hour downloading and installing it just so they can sign the useless agreement.

One could make the argument that anyone speaking at LinuxWorld should have OpenOffice installed on their machine. After all, LinuxWorld is an open source event...right?

But this argument is weak. No one should have to use open source. If open source can only compete by edict - government, corporate, or otherwise - it's not worth having. At Alfresco, for example, we believe we've created an Enterprise Content Management System that is demonstrably better than its proprietary counterparts - in features, robustness, ease of use, and cost. Open source is an efficient way to develop Alfresco, and to sell it. But it's not a crutch ("Hey! The product stinks, but at least it's free/open source!").

All that said, I guess fair is fair. Microsoft imposed its formats on us forever. I guess it's fair to have the open world impose itself on us, too. Annoying, but fair.

I signed up for both partner programs, and all I got was this lousy...

I think Novell has been more generous to me in my departure than they were while I was an employee. I signed Alfresco up for Novell's partner program in December, and just got this in reply:

Congratulations! The Novell Partner marketing team ran a promotion for companies that listed their products in the Novell Partner Product Guide in December, and you have won an Apple iPod Shuffle. Please give me an address and phone number for where you want the iPod shipped, and we'll send it out to you.

Thanks for your participation in the Novell PartnerNet for Technology Partners program.

Novell has proven to be a responsive, generous partner to Alfresco and many others in the open source business community. I'm grateful for the iPod, but I'm much more interested in the efforts Novell is making to build an open source ecosystem around SUSE Linux. Its MarketStart program is just the start....

Funny enough, I also got an email today from Red Hat. This one came from Matthew Szulik. Here's a blurb (which I quote only because the email isn't intended to be a private communication):

2005 was an incredibly important year in the evolution of open source technology as a catalyst for change. Major technical accomplishments were made and important political milestones were reached. Firefox marked its arrival in the mainstream with resounding industry accolades. The EU ruled against Microsoft in favor of innovation. Open source is making technology accessible to the masses through the revolutionary One Laptop Per Child project.

As we look forward to 2006, I feel that we are at the vanguard of a new revolution in the potential of open source as an economic model and its impact on the marketplace....Red Hat has invited a group of the leading open source leaders in the industry (commercial & non-profit) for a half-day “meeting of the minds” to discuss trends, issues and opportunity for collaboration. I truly hope that you are able join us for this meeting....

The meeting will provide a forum for open discussion on commercial open source as well as community and foundation projects . The goal is to design larger market opportunities for all organizations that create products and services based on open source technologies. In addition, members of Red Hat's executive management team will outline our technical and business roadmaps, and solicit your feedback on our direction and role in the industry.

Red Hat has finally started taking its leadership position seriously (and OSBC's, given that they planned this skull session to directly follow OSBC San Francisco next month), and (apparently) realizes that it has an obligation (and an opportunity) to build a vibrant open source ecosystem around itself. Those new to working with Red Hat won't realize this, but this is a dramatic shift in mindset for the company. A very welcome and productive one.

Now if I could just get Red Hat to give me an iPod Shuffle. :-)

The changing tide in technology

I had an interesting conversation today with one of Alfresco's system integrator partners today. The company in question has been a long-standing Documentum/FileNet/etc. partner, and has built a thriving practice in partnership with these big, proprietary companies.

I assumed, therefore, that her primary interest in Alfresco was her familiarity with our founders (one of which, John Newton, founded Documentum and worked with her there before they each left for other opportunities). I was wrong.

It turns out that she could sense a change in the dynamics of the industry, well before open source made even the remotest imprint on her territory. She didn't know much about open source when she contacted Alfresco. All she knew was that the dynamics of the traditional software business were souring:

• Enterprises were tired of proprietary vendors promising features - even demoing them - only to leave the SI to implement the vaporware (and assume the costs of developing what wasn't there);
  
  
• Enterprises were sick of putting out huge sums of money - way too much money - up-front for software that, more often than not, failed to live up to its marketing; and
  
  
• Enterprises were annoyed at the distant relationship they had with their vendors, who only cared about them long enough to make the sale.

So, this particular SI started casting about for something different, something that reflected a solution to this buyer fatigue. And she found it in open source. By accident (in that she was initially more interested in seeing what John was up to than in open source, per se), but it was there, all the same.

I've been watching this shift since 1998. But today was the first time that I really felt the shift from someone who has "skin" in the proprietary game, but sees that game quickly running its course.

Open source, quite simply, is a better way to reach customers. To sell them on substance, not chicanery. To lower the cost of getting business needs met, without pilfering from vendors' or integrators' pockets. (There's more than enough money to be made in smart, innovative open source software. The business models are there.) It's a better way to build a software business. Period.

A very happy new year for MySQL

In my in-box this morning was an update from Marten Mickos detailing the company's progress over the past year. Marten is one of my favorite people, and MySQL is one of my favorite companies, so it's great to see them doing so well. There are people (or soccer/football clubs like !%!@%!% Chelsea) that are so obnoxious in their success that it's hard to feel anything except extreme animus toward them. Not so with Marten and MySQL. They deserve their success, and it's easy to be glad for them.

As it was a public email, I'll post it here.

Brief MySQL Update

In 2005 we continued to make free and open source software available and affordable to all. Version 5 of the MySQL server became an instant success with over 1 million downloads in just three weeks, taking the MySQL active installed base to 8 million. Millions of hobbyists, academics, open source projects, start-ups, software vendors, large corporations and government agencies are building new applications that earlier were not economically or technically feasible. In June the EU
parliament made an important contribution to the health of the software industry by rejecting the so called software patent directive.

As a company we grew from less than 200 to nearly 300 employees in 2005. We strengthened the management team with an EVP Sales and a CFO. We set up operations in Japan, and we signed up powerful distribution channels for MySQL Network, the subscription offering we launched early in the year. MySQL Cluster proved itself in the market for high availability databases.

In Q4 of 2005 it seemed that the entire database market was ready to bet on open source. We saw an acquisition, a spin-out, some major new releases, zero-priced limited product editions, and more. What a wonderful public validation of our business! Sales-wise we rounded up the year over target, with our best month ever, best quarter ever, and best year ever.


What did we learn in 2005?

Open source is maturing in a healthy way. Apache, Linux, PHP, Eclipse, JBoss, Firefox, MySQL and a few more act as locomotives pulling new projects and companies with them. And venture capitalists are increasingly eager to help them on the way.

The LAMP stack is becoming a serious contender thanks to a host of new companies that fill out the functionality gaps (SugarCRM, Alfresco, JasperSoft, Asterisk, Zimbra, ActiveGrid, to name just a few). Vice versa, start-ups can quickly become serious contenders by building their solution on the LAMP stack (MySpace, Flickr, Facebook, Neopets, again to name just a few).

It seems that the most phenomenal financial performance is achieved by companies who deliver their service online and who build their infrastructure on open source. I am thinking of RightNow, Google, Yahoo, online gaming companies, and others.

For a software vendor with a growth ambition, closed source may not be an attractive avenue any longer. But neither is a support-only model on open source. It appears that the most resilient and growth-oriented model is one that combines true open source with a compelling commercial differentiation.

[MATT NOTE: Bingo! Marten is dead-on with this insight. Support is not a great business model, but support + Marten's "compelling commercial differentiation" is. This is what SugarCRM does, what Alfresco does, and what an increasing number of open source startups are figuring out.]

Gartner expects 75% of IT organisations to have an open source software acquisition & management strategy. Organisations have a 12-18 month window to get their strategy in shape and those that don't will be at a competitive disadvantage.


What’s awaiting us in 2006?

It seems that buyers will increasingly prefer to buy software and support combined in a subscription offering rather than having to pay upfront licence fees. Corporations will rejuvenate their IT infrastructures by adding open source scale-out components to existing systems. And I believe that open source stacks will evolve faster than closed ones, in line with the teachings of both Charles Darwin and Adam Smith.

Software patents and liability issues around software licensing will stay on the agenda of both buyers and sellers of software, but perhaps in 2006 it will become a manageable issue with a set of best practices.

I hope that vendors will finally agree that commoditisation is a higher, not lower, level of innovation – offering more opportunities than threats. Red Hat’s impressive growth numbers and the new strategy of Sun Microsystems support this view.

Closed source software vendors may take more drastic actions to join or to fight the open source movement. And who knows - more DBMS vendors may go open source.

Finally, there will be a growing group of young software professionals not knowing that there was a time in the 20th century when software was not open and free. Perhaps William Gibson thought of them when he said “The future is here, it’s just not widely distributed yet.”

Thanks for all the support in 2005 and enjoy the year of 2006! Please, as you undoubtedly will, take my predictions with a grain of salt.

Marten Mickos

P.S. I hope to see you at the 2006 MySQL Users Conference – www.mysqluc.com.

Keep it up, Marten. You exemplify the best open source has to offer: aggression, humility, collaboration, affability, and competition.

Safety in open access

Having just finished William Tyndale's biography, I suppose I'm in an ecclesiastical mood. (Though I'm balancing this mood out with Thomas Hardy's Jude the Obscure. Moderation in all things, you know. :-)

What better, then, to read than Eusebius' Ecclesiastical History?. I've had it for some time, using it occasionally as a reference, but this morning was the first time I've actually cracked it open and started reading it, front to back. Roughly 100 pages into it, I came across an interesting chapter detailing which books/epistles/etc. are counted as scripture, and which are not.

The fascinating part of this is the role that authority plays in making such decisions. For instance, I've read a fair amount in the Old Testament Pseudepigrapha, as well as the New Testament Apocyrpha, but these are not considered "official" books of scripture. Centuries ago, a small group of experts determined what was official, and what was not. (An example of ongoing debate on the matter can be found here.)

I appreciate the role of experts as well as the next person, but I'd rather trust in numbers, in open access. This is, I suppose, a Wisdom of Crowds-type argument. But really it's more a testament to my personal faith in openness. I think rubbish generally outs itself over time, as does truth, scientific or otherwise. Give enough people access, and good things will prove themselves good, and bad will prove themselves bad.

This is not dissimilar from William James' Pragmatist philosophy - whatever works, when put to adequate tests, is "true," with the caveat that it may be proven untrue down the road after further experience. As Karl Popper writes, however, an open society is necessary to achieve the optimal forum in which to determine what works, and what doesn't.

Access is critical, both to those who look and to those who don't. As I've written before, the option of source code access in open source serves as a useful surrogate for actual exercise of that choice, just as an open process for choosing scripture over apocrypha, open voting, open/free market economies, etc. provide for (or, at worst, enable) better informed, optimal results.

No decision is best made blindly. No product is best defined, designed, and implemented in an information/feedback vacuum. Opening up source code means customers can place greater trust in the software they use even if they never read a single line of code, precisely because others can exercise this choice in their stead.

I need to spend more time thinking about how to implement this openness in one's business model. I'm hoping that you, along with I, will spend a great deal of time working this out in 2006.