Milton Keynes famously said, "In the long run, we're all dead." Keynes was trying to focus economics on the short run - the world that matters here and now, and not expend unnecessary brain cells on what might happen well down the road. Largely a good principle.
It might also apply to the Long Tail I blogged about the other day. As Gaping Void humorously points out, success may well be a "short tail" phenomenon.
So, I can't claim credit for the intelligence in this post. I'm posting an off-line comment someone sent me, without attribution (because I assume they'd post it publicly if they wanted attribution). Great comments. The commentator has an excellent background in the space, which reveals itself in their analysis.
Firstly, I'd observe that the world in which applications consume data ==revenue for the MNOs [Mobile Network Operators] is an interesting one, but there's a number of weak-patches in it which MNOs don't want to acknowledge. From the competition for $/bit (WiFi etc) through to behaviour encouraged in developers (please, consume lots of data!), through more crucially to the security issues (paying for data leaves a whole new level of exposure to malware, that's much more subtle to detect), the space is fairly bleak. One of the interesting data points here is actually how the MNO's view this space - they all see it as having a small number of applications (that they themselves write) driving the lion's share of the revenue, rather than a vibrant, open, development community.
Secondly, the more interesting aspect to the problem is how silo'd the space is getting - the struggle between MNO's to 'own' the customers vs. the handset manufacturers getting commoditised (witness Nokia's profit margins) is locking down devices rather than creating an open interoperable community. To me the real opportunity lies where I can take the high-bandwidth, high processing devices (servers, PCs) and crank on them to distill out appropriate information to my low-bandwidth, low CPU mobile device, and reciprocate with temporally relevant data updates.
Finally, the development challenges in this space are a real hurdle, because the mobile space is just simply following in the PC's wake. They view the PC's (Windows, Linux, Mac) as development nirvana, and as a result are letting mobile devices in for even worse issues than in the PC space as it struggled to accommodate ubiquitous connectivity. Mobile handsets are by definition connected, and even more so when bluetooth takes hold further, and as a result the prospect for damage without the right security model in place (not borrowed from the PC world) could be incalculable.
Joe Kraus has a fascinating blog entry today on the long tail of the Internet. It coincides with separate self-introspection as I've been thinking through my own blog. I get a fair number of people reading this blog each day, which (frankly) shocks me. I never promoted it in any way, and yet people find it, just as I find new sites/information on a daily basis. This blog is part of the Internet's long tail....
I'm not Google. I'm not Amazon. Not even eBay (though I do have a lot of junk at my house). And yet there's a comparatively small group of people who think that something on this blog is worth reading occasionally. (My children are not among this number - they rightly find me dull.)
Frankly, open source is also part of this "long tail" trend. Developers can write and distribute code without a gargantuan target audience in mind. It might just be me, or some other developer, who finds it and finds it useful. The Internet facilitates this through search (including SourceForge) and peer-to-peer. I've criticized this before, arguing that most open source usage bulks up around very few projects (Linux, Apache, etc.). I'm not sure why this is a bad thing anymore....
Millions of markets of dozens, as Kraus dubs the phenomenon. Peer-to-peer open source. If it's peer-to-peers, all the better. But one seems to be enough.
At least, to get the code generated. Is it enough to sustain a business? I don't know yet, but I hope to have it come up at OSBC and other gatherings of people interested in turning open source into viable business models.
The votes are in, and this is the list of startups who will be showcased at the Open Source Business Conference next week. If your company is not on the list, let me know why we should have had you there. We're doing another event in Boston (October 2005) - we'd love to see you there. That event will be heavily targeting early-stage companies and strategies - kind of like an open source DEMO.
Realm Systems
Optaros
Zend
Covalent
SpikeSource
Forum Systems
Funambol
SugarCRM
Univa
Humano2
OpenLogic
Simula
Coridan
ActiveGrid
Compiere
Open Country
OpenXource
Protique
Mergere
Digibug
Cymphonix
SourceTap
Active Endpoints
SafeDesk Solutions
BlueSunrise
Antigrain
BitRock
Groundwork IT
Adisoft
Aduva
Palamida
Sleepycat
SocialText
Transitive
Dave is a good friend of mine, and a fellow commiserator on conferences. He does Linuxworld; I do OSBC. He has a great collection of rants on why submissons aren't accepted for Linuxworld, which you might find interesting, especially if you're planning to submit a proposal.
I've long admired Soft*Letter's analysis of software companies. Soft*Letter does a great job of tracking how much software companies spend on sales and marketing, etc. It was while reading Soft*Letter's analysis of sales and marketing spend that I had the ah-hah moment about open source distribution (versus development) that has guided (plagued ;-) my thinking ever since.
I'm reading new material from Soft*Letter, and am seeing my Dilbertian life spelled out in black and white. Software companies (according to Soft*Letter) spend 20-22% of revenues on sales and marketing. When these companies really strike it rich and go public, they end up spending 45% (median).
Why? Well, in part it's because Soft*Letter's 45% number came in the middle of the downturn, when companies were struggling to increase revenues. But the primary reason, I suspect, is because Wall Street demands (unnatural) growth, which causes somewhat unnatural spending. Throw more bodies at a problem and it will be resolved faster. Well, this isn't true in engineering (as classically captured in The Mythical Man-Month, and it really isn't true elsewhere. Productivity is largely an individual effort, and individuals are capable of doing much more than big companies normally give them credit for.
Other interesting points, with my commentary:
Points of Resistance: In some market segments, high levels of S&M spending seem to be necessary for almost all competitors. In particular, the enterprise and education markets continue to require unusually high outlays to overcome customer resistance; by comparison, vertical market buyers seem to be much less costly to reach and persuade.
Sales channels: Companies that rely on indirect channels - primarily VARs and retail outlets - tend to have substantially lower S&M costs. Those with a direct sales model often spend twice as much as their indirect counterparts for every dollar of revenue they bring in.MNA Comment: This paragraph screams out for the need for open source. Enough said.
A thought occurred to me while reading a Forrester report on the mobile telecom industry today. Though many in the US may find this surprising (because we're at a 57% penetration rate for mobile phone usage), Europe and APAC are at 90%+ penetration rates, in many cases. For example:
United Kingdom: 91.6%
Italy: 98.4%
Germany: 78.8%
United States: 57.2%
Japan: 65.0%
China: 22.2% (but growing exponentially)
Taiwan: 103.5% (ever the over-achievers ;-)
The voice markets for these countries, then, are largely saturated, driving wireless carriers to look to content for increased revenue per subscriber.
Easy enough, right?
Wrong. No one seems to be able to guess which services will take off (SMS) and which won't (watching TV on your cell phone - who wants to do that?). However, one thing is clear: the revenue model affords software companies different, exciting opportunities.
What do I mean? Well, consider that the wireless carrier is most concerned with bumping up data traffic. If your application brings that traffic, they should (and, as I'm learning, will) pay you for it. So, instead of monetizing customers by charging per-user license fees (traditional software model) or through advertising (Google/etc. revenue model), you make money simply by driving traffic.
Think about the value this provides to software creators (especially the open source world): suddenly, your primary concern is optimizing and maximizing use of your software, rather than finding ways to goad would-be buyers into paying. For that matter, consider the value to consumers/users: "free" software. (Of course, you pay for it every time you pay your wireless bill, but it's indirect.) I'd actually published a related article nearly a year ago, but I was talking about the desktop then - I think the mobile phone offers a better platform for the indirect usage model.
The question arises, of course, as to whether or not wireless carriers will enter into contracts to pay software creators based on data usage. I think they will. I've already seen nascent examples of this happening. Interestingly, the P2P networks that supposedly decimate revenues and profits for record labels and movie studios would be a boon to wireless carriers, provided they enter into agreements with the networks to drive that data traffic. If they don't, the P2P networks will end up clogging their networks without them capturing a dime (but paying plenty).
Still plenty of questions to be answered as to the model's viability, but this is a promising area for analysis. While Booz Allen Hamilton's assessment may be true that mobile subscribers won't want to pay for their data usage (though I do this happily on a monthly basis), I think their concern is overblown. Subscribers may opt for all-you-can-eat data/voice plans to escape the vagaries of a pay-for-data plan, this is no different from what we already do.
In most cases no other OS was considered. Linux was the choice from the start of the product or even company itself. The real reason comes down to cost, however the infrastructure or ecosystem around Linux enables to the decision to be made. That is, the tools, open source components/commercial middleware and availability of skilled knowledgeable labor are sufficient in the minds of the decision makers to allow a cost based decision. Linux is just "good enough" in many cases.
Call it a fetish. Call it whatever. I really like the embedded market, especially the mobile sub-market. There's something about helping to build devices that normal people use every day.
I had lunch (and a jog - let's just say that Ken is in much better shape than I) with Ken Klein, CEO of Wind River, on Monday. I managed to gasp out a few questions, and really liked Ken's answers. This is a company (and management team) worth betting on.
Let's take a look as to why.
First, the supposed reasons that the embedded market stinks. The embedded software market is relatively small, highly fragmented, and has been technology's most ripe market for open source, especially Linux. That's right: Linux is big (huge) in embedded, but that hasn't been a great thing for embedded Linux vendors (like my former employer, Lineo, which Motorola bought in 2002). As much as 80% of embedded development is RYO (roll-your-own). Now, instead of rolling their own operating systems, vendors have just been rolling their own flavor of Linux.
There are no clear leaders in embedded, beyond Wind River (and MontaVista in embedded Linux), and WIND has faltered in the past few years (largely due to the telecom bust, which had been its biggest market, but also because of the general market decline). Microsoft has struggled to maintain a beachhead. Red Hat bought Cygnus and then promptly forgot about embedded - I think they likely pull together $5Mish at most from the market (and have never really competed in it - we never saw them in deals at Lineo, and I know from talking with friends still in the industry that they remain sidelined). In sum, no large vendor has yet assumed technological and business leadership of the embedded Linux market.
This fragmentation has contributed toward keeping the (commercial) market relatively small. Venture Development Corp. (VDC, the premier analyst firm for the embedded market) estimates the embedded OS market at $762.1 million in 2003, growing to $1.1 billion by 2006. Adding in tools and related services shows a total market size of $1.4 billion in 2003, mushrooming to $2.07 billion by 2006. (The embedded Linux market? $65.2 million. Not something to cheer about.)
Of this market, the two primary submarkets are Consumer Electronics and Telecom/Datacom, with Industrial Automation coming in a distant, but significant, third. Historically, Wind River's VxWorks and pSOS OSes have dominated the latter two markets, with a range of companies (Microsoft, Palm, Symbian, etc.) competing for the Consumer Electronics market.
When the biggest player in the market (Wind River) is only making a few hundred million (sub-$300M), that's not necessarily a healthy market.
Right?
Not so fast. That was then, this is now. Importantly, the $762 million market figure may only be the tip of a much larger potential market. Given that typical embedded applications have grown from 100,000 to 1,000,000 lines of code in the past two years, and that OEM lead times have dropped to as little as six months, in-house developers are increasingly incapable of keeping up with project requirements through roll-your-own infrastructure.
This is a great boon to embedded software vendors. The commercial embedded market has been constrained primarily by stubborn resistance from the roll-your-own market: as much as 80% to 90% of prospective embedded buyers choose to build their own operating systems (or to customize Linux/Debian/other). While there is money to be made in taking customers from proprietary OS vendors and from competing embedded Linux vendors, the real money is to be made in converting roll-your-own customers to off-the-shelf customers. The market is demanding it.
With this in mind, the embedded pie could balloon to anywhere from $3 billion to $20 billion annually as these in-house groups look to commercial products (depending on how far one believes that all of that RYO market could or would convert to commercial). This estimate defines embedded software as including other components, such as middleware, services and new layers of applications that lie somewhere between the OS and the end application. It makes for a big market, and one that needs vendors more than perhaps it ever has.
In short, I believe that the embedded market is ripe to explode. Who will capitalize? Those like Wind River who provide end-to-end solutions that get customers (ODMs, OEMs) to market quickly, at a reasonable cost, while also allowing flexibility to differentiate products. Open source vendors can compete here, and can do so effectively, but they need to keep in mind that free software really doesn't get you far in embedded - not with a six-month product development cycle. Customers need to see compelling value to save them from rolling their own.
MyFamily.com is nearing two million paid subscriptions. Subscriptions range from $7.95 to over $230 a month. The company far exceeds $100M (I'm guessing that they're nearing $150-200M) in revenues, and is immensely profitable. For those who think it's a "Mormon thing," consider that most of its users come from outside Utah, aren't Mormon, and the management team is overwhelmingly not Mormon. It's not a Mormon thing.
It's a money thing. There's gold in them thar genealogy hills. Genealogy.com reports that 80 million Americans are currently engaged in tracing their family roots, and is the second-largest hobby for Americans today. (I guess this puts it behind guzzling beer while watching March Madness and the SuperBowl. Who knew? :-)
Genealogy is fun. It's really exciting to figure out, jigsaw puzzle style, which of your ancestors were convicts (all the while sticking to your comfortable view that you come from nobility). It's also unnaturally difficult today. Why? Because the search technology used by MyFamily.com and the LDS ("Mormon") church stinks. It's pathetic. I can find my way around the Ancestry.com site (the site that MyFamily.com runs) because I know how to decipher census records, immigration lists, etc. But then, I've been doing this for 10 years.
The problem to be solved - searching and delivering relevant results on millions/billions of names/dates/relationships/etc. - is one that Google (or Yahoo! or Microsoft) is well-positioned to solve. The data is (mostly) public domain. It just requires someone to search it. For that matter, someone could take Nutch, the open source search engine, and apply it to the problem.
Like data warehousing queries today, genealogical searches on Ancestry.com require structured queries, rather than Goolgesque searches. You have to know where to search, and you have a very limited range of what you can search for. This keeps a vast majority out of the market or, at worst, keeps the existing market from benefiting from the service fully.
This is a big-money opportunity. And it could be completely open source, in every aspect of the name. All of the data could be open, in Amazon, eBay, Google, etc. style. It's the service that generates revenues - what you do to the data. Apply sophisticated search technology; wiki-fy it, by pulling family trees into the system (by default, in Clay Shirky/Tim O'Reilly "architecture of participation" fashion) and linking them together to create a meta-tree; even patch it into Flickr and other photo sites such that the flat-file names can take on personality through pictures; etc. It's a great opportunity.
Enough on this. I was just thinking about it this morning because I recently did a "quick" genealogy search for two friends. It took several hours to accomplish something that a Google/Nutch service could have completed in a matter of seconds. The technology is out there - it just needs to be applied.
So, maybe I'm wrong. I've always felt that open source needs a marketing team to put the last-mile polish on it to make it useful for customers. Geeks develop for other geeks (so the saying goes), and aren't necessarily in tune with the needs/requirements of the end-user.
But in a conversation I had with Fabrizio Capobianco, CEO of Funambol (Means "tightrope walker," apparently - metaphorical for the company's attempt to balance open source with commercial interests), I had this view challenged. Fabrizio is a sharp guy (his affinity for Juventus being just one sign). He told me that unlike standard development, open source development gives his development team front-line access to customer complaints/feature requests/etc. Instead of sitting in a tranquil DMZ, shielded (and informed) by product marketing, his developers actually hear real-world customer feedback all day long....
So, perhaps open source is more customer responsive?
One more, and then it's off to bed. (Arsenal match on at 5:30 AM tomorrow - can't miss Thierry Henry in action.)
Driving home from work the other day, I was remembering one of the best theatrical productions I've ever seen. When my wife and I lived in the UK, we went to see an old favorite, Guys and Dolls, at the National Theatre in London. It's one of those shows that every community theatre performs at least once per year (as well as Joseph and the Amazing Technicolor Dreamcoat, if you live in Utah, where there always at least five theatres staging it at any one time). A comfortable old favorite.
This production, however, turned out to be amazing. Same musical score. Same lines. But the actors/singers were unbelievable. When the actor playing Nicely Nicely Johnson launched into his rendition of "Sit Down! You're Rocking the Boat," the crowd went crazy. It was the most electric performance I've ever seen, hands down. (Metallica doing "Creeping Death" comes close, but in a very different way.) Encore after encore. I was crying through it all, it was so exceptional. I left the theatre that night feeling glad to have been alive to see such a performance.
What does this have to do with software? Nothing.
Well, something. Just as I was happy to pay for excellent wrapping around a commodity script/product, so, too, is it possible (or, rather, likely) that IT buyers will gladly pay for exceptional wrapping around commodity (by which I mostly mean "standardized" or "standards-based") software/hardware. We love TiVo, despite the fact that it's just Linux in a box, with a super-intuitive UI. Google? Ditto. Palm OS? Yep. Same thing.
It turns out that for all our bleating to the "innovation" tune, the best innovation is really that which makes complex products work simply. Apple is the master of this, but they're not completely alone. Just lonely.
I wasn't hot on the prospects for open source's rising breed of system integrators like SpikeSource. I'm changing my tune, though, because of Guys and Dolls. Reflecting a bit more on my past posts about innovation and commoditization, I realize that what I really want is for open source companies to put pretty packaging on the vulgar complexities of unpolished code. Red Hat did this first, though I believe Novell is actually doing it better (and soon-to-be released IDC numbers will confirm that customers agree).
But my employer aside, this is green fields territory for open source startups. IT users need mind-numbingly simple products to buy. They don't necessarily care if these are open or closed source in nature (though I think it's foolish to build on a closed platform). They just want it to work, and to be intuitive plus cost effective. By leveraging an existing code base, open source startups are privileged to dedicate their development cycles to polish open code to a Guys and Dolls shine. Like Apple's OS X and Safari. Like FireFox. Etc.
So, Sister Cooper finally left Sun. I don't mean "finally" in any derogatory way (toward Sun or Danese). I just think it's about time. It's good for her to expand beyond Sun, to enable her to extend her reach beyond one organization, and give further proof that Danese is a community player. I can yap about whether or not Michael is overly partisan, but if I ever suspected Danese was, the move to Intel dissipates it. [Btw, here's her blog.
Frankly, I think it's about time she left Corporate America altogether and did the Larry Lessig thing, lobbying for the open source community full-time (unhindered by corporate policies - not that these have had a measurable effect on her opinions ;-). But there's probably some truth to the notion that she can do more good within the walls of a company than outside. She has street cred within the open source community - working for Sun or Intel gives her corporate cred, and that's important (to the group she's trying to convince). And a paycheck is nice. Last time I checked, Intel was still paying those.
Some have pooh-poohed the move, and Danese's relevance. I doubt I would have stumbled across it (like porn and other crap, I don't seek out destructive time wasters), but Danese pointed me to a rant from some Unix bigot, trashing her. (I won't link to it here, because I have no interest in pointing people to rubbish.) That's like her: perhaps miffed at this person's misinformation, but happy to take a few bullets. She does that a lot for the community.
At first, I really didn't know what to make of her. I'm fairly conservative (see top left description). Danese wears funky glasses, hangs out with the developer crowd in a way that I can only dream of (understanding jokes that begin with C++ and end in "Hello World"). I also didn't know what to make of the crusader-for-open-source role she seemed to play.
But it's legit. I've never seen her do or say anything that comprises her (rightly given) title of "open source diva." Whether self-annointed or community-appointed, it fits.
Given all this, the Intel news is a bit of a non-issue. Who cares? Danese is Danese whether taking a paycheck for Intel, Sun, or whomever. We're lucky to have her.
I spent some time talking with Dave Nagel, CEO of PalmSource, today. In the course of our conversation, he said something that I found very interesting. He mentioned that Linux-based phones are predominately a phenomenon of the high end of the mobile phone market. RTOSes (real-time operating systems) like Nucleus and Vertex dominate the vast majority of the current mobile phone market, with Symbian also taking a chunk. These are the low-end phones/feature phones that the vast majority of the world uses.
But the emerging market for high-end smart phones? That's going to be a Linux world. Some Windows Mobile. Some Symbian. But lots and lots of Linux.
"Linux" and "High-end." The two are not normally thought of as synonymous, at least not in the enterprise world. There, Linux servers nibbled at edge servers (file and print, email, etc.) before starting to devour the data center (which is where we stand today). Through high-performance computing and such, Linux is becoming high-end in the enterprise world. But that's where it begins in mobile.
This is all the more interesting because, as is becoming clear to me, the mobile world is perhaps the least commoditized of computing markets. Standardization has not set in there the way it has in desktops and servers, leaving a great deal of room for full-stack integration (a la Clay Christensen's "innovator's dilemma" theory). Clearly, Linux in this world is a choice based on its viability/stability/community nature, and not its cost (the RTOSes of the world are cheap to free).
Linux is high-end in mobile, in other words, because it's a high-quality, best choice solution. Not because it's a commodity. Not because it's free. Nokia (Symbian) and Microsoft suddenly have a big problem on their hands. How can they compete with a better performing and better perceived OS, which happens to be community property, and extremely low cost?
(Hint: They can't.)
Jeff Bezos (CEO of Amazon.com) is here at eTech, and is walking the halls, and attending sessions. (Just bumped into me.) It's pretty cool to see a CEO of a major company mingling with we commoners. I actually think he cares about the various geeky sessions, and is learning from them. How refreshing.
I'm at O'Reilly's Emerging Technology Conference, listening to Tim expostulate on the new rules of building successful apps. First off, we need to stop thinking of apps as icons on our desktop, and we need to stop thinking of the web as a series of URLs. Google is an app. Amazon is an app. Etc.
Jason Matusow has a hard time understanding "open" source when viewed through the lens of the GPL. Jason views the BSD (because it is more permissive) as more open than the GPL, showing that he hasn't yet fully bought into open source as defined by Richard Stallman. Stallman, of course, hates the term "open source," considering it an abomination (to put it mildly). For these, a license (and its software) is not truly open if the possibility exists for co-opting it to take it private.
Hence, for the "free source" devotees, the more restrictive the license (in guarding the liberties associated with free and open source code), the more open it is.
It is not surprising that Jason would have difficulty comprehending this (or, at least, buying into it). The country he grew up in, for example, explicitly disavows this type of freedom. Freedom that must be forced is not freedom. Instead, forcing people to enjoy freedom smacks of tyranny. Just as we would not consider ourselves free as citizens if the government watched over us, AK-47s in hand, to ensure we voted, wrote letters-to-the-editor, and otherwise lived up to our freedoms, so, too, it's not really free software if the license paternalistically forces us to live up to someone's ideology.
This is not to say that such code creators are outside their rights to do so. It is only to challenge the notion that their way enables more freedom. It does not.
It is a bit like a debate I had during my Masters program with Stefan Rossbach, a professor of mine with a fondness for Machiavelli. (In is defense, should he want it, I think Stefan was playing this card more for effect than anything else.) Stefan argued that Machiavelli was actually one of the most moral of political philosophers, because instead of leaving outcomes to chance, Machiavelli insisted on a political regime that took complete responsibility for the results of its policies. This sounds fine on its surface - surely, everyone should be held accountable for their actions.
Well, the only way to guarantee results is through force. Through control. Through an authoritarian regime. In American democracy (at least, in its ideal form), we expect citizens to learn the law and live according to it. To the extent that people live outside the law (generally measured by when one's actions negatively impact someone else), the law punishes. It is a benevolent cycle, despite Michel Foucault's silly Discipline and Punish.
Similar to Rossbach's Machiavelli, Stallman and others in the free source camp have trained the emerging open source community to believe that only through force can true freedom abide. They are wrong. They are objectively wrong (Apache and other projects governed by permissive licenses have not sold out to "The Man" any more than Linux and its ilk). They are subjectively wrong.
In line with my subjective view is my moral view (already evident above), which is that no matter how much we may want everyone to freely share code (in GPL-like fashion), we should not achieve that end through force.
Of course, this argument is tempered somewhat by the fact that we're not dealing with essential freedoms here (despite what Stallman may think, software licensing really has a somewhat negligible influence on the lives of 99.999% of the world's five billion-plus people). In addition, if I don't like a certain license, no one is forcing me to use it. I only have to worry about a given license to the extent that I want to leverage that software. I'm always free to use other software, or to write my own.
All true. My only complaint is, with Jason, that we mischaracterize open source. Apache, because of its license, really is more open than Linux, because Apache achieves its end through choice, rather than force. This distinction matters.
All this being said, I still prefer the GPL as a competitive license. It is a fantastic distribution mechanism, and an excellent competitive club. It's also an excellent way to have competitors collaborate without collusion (illegal) - since no one can effectively own the common property, it provides a safe place to create communal property (like Linux). You just don't get that same effect through a BSD-style license; or, rather, you don't have any legal guarantees that it will turn out as such. So, the GPL is a valuable tool - just don't call it the most open of licenses.
First, a personal anecdote.
I spent Saturday driving down to San Diego with my wife and kids (to the O'Reilly Emerging Technology Conference). On our drive down, my 8-year old daughter (an animal lover) said, entranced by all of the bovine animal life visible out the windows,
"Look at all of the cows!"To this, my less-than-vegetarian 5-year old son said,
"Yeah, that's a million-billion hamburgers!"
...that Jason Matusow was a music major (saxophone) when he was at Boston University? I had always pictured him as a hockey player. Like ogres, I guess Matusows, too, have layers. :-)
I hereby demand that Jason perform a solo piece at OSCON this year. Make him commit, Tim!
I get frustrated with the tenor and depth of discussion in the technology world sometimes, especially as it relates to open source. We're inundated with sophistry (fairly interesting, but vapid) and name-calling (not interesting, and vapid), to the point that we never really clear the haze long enough to have intelligible conversations about how to solve customer problems. Waxing Biblical for a moment (r0ml does it, so I can, too :-), we "make a man an offender for a word" (Isaiah 29:21), to the point that people are afraid to say anything of note.
The phenomenon is mostly a prisoners' dilemma of sorts. In this game, the vendor community views the world in zero-sum terms, and capitalizes on the slightest (perceived) missteps by competitors. No one wins in this situation. Rather than engage in healthy, open dialogue, we hide behind buzzwords, backchannel half-truths about competitors to our customers, angled TCO studies, and such. We end up with a load of rubbish to guide customer decisions. I doubt vendors get much better marks than politicians in surveys on vendor integrity and trustworthiness.
Here's a case in point. Jason Matusow, Microsoft's Director of its Shared Source Initiative (and a wonderful person, though if you've met his wife as I did tonight, you'll readily admit that his wife outshines him - Jason definitely married above himself. Sorry, Jason), has just kicked off a blog of his own. Those who have heard Jason speak know that he's exceptionally sharp and a great credit to Microsoft. Sophistic about Microsoft's product strategy at times, but that's what he gets paid to do. I've never heard him bad-mouth open source, but instead offers erudite opinions as to why Microsoft's strategy is better (and, where it's not, how they expect to adopt open source development methodologies). Right or wrong, it's an intelligent position. Spend more than a nano-second talking with Microsoft, and you realize that the company has a more nuanced perspective on open source than normally credited.
And yet I doubt it's an opinion he'll be able to craft in print, because the slightest misstep will be analyzed and reported to death. As such, it's likely that Microsoft PR will lobotomize the more interesting things Jason might feel inclined to write. I hope not, but in the business climate we work in, I'm not overly optimistic.
I think it's time that we, as an IT community, hold back a little and provide more room for our various competitors to make their arguments/pitches. We need to move past binaries: Microsoft = Bad, Open Source = Good. I can think of many instances where I think Open Source = Better Choice and Microsoft = Worse Choice, but these are data-driven decisions/choices, not something I've opted for because it fits my desired outcome/ideology. Use whichever technology/product works best in a given situation. (Gartner, Forrester, IDC and others can help with this by not allowing their research to be skew-for-hire.)
In short, we need to move beyond ideology. Open source does not need ideology to win. Not anymore, anyway. Ideology may be important in the early stages of a movement. You need people to advance a cause in the face of tough odds, and this may well require a somewhat unreasonable imperviousness to criticism and contradiction.
But we're beyond this stage now with open source, or should be. It's an accepted development methodology, with an increasing number of poster-child projects to hold up as proof of its success. We can move past the ideology, and offer up reasoned, introspective criticism without feeling that we're somehow "apostatizing" from open source "faith." Imperfect open source will still be better than imperfect Microsoft or Oracle or [Insert company name here] in a wide number of scenarios. In all? Likely not, but we can still, turning Emily Dickinson's words against her, tell all the truth without telling it "slant."
I'd like that. Customers would, too.
So, Stephen Walli is a good friend. Stephen understands open source, and is a great advocate for open source. He pushed open source at Microsoft (succeeding in getting a few projects floated on SourceForge), and is now at Optaros. They're lucky to have him.
That said, I still can't stop laughing whenever I think about the Optaros' announcement of Stephe's joining the company. His value to the company is clear, but the phrasing of the announcement gets me every time.
OPTAROS NEWSIs it just me, or does it sound really funny to be looking to Microsoft for understanding of open source development communities? (I know, I know, Stephe! But it's funny how it's worded, all the same.)
02/18/2005
Optaros Hires Ex-Microsoft Manager to Build Relations With Key Open Source Software Development Communities
I had always appreciated the open source world's non-confrontational manner on my desktop. That is, open source programs waited patiently for me to use them, and never tried to impose themselves on me.
No more, apparently. I use OpenOffice (actually, NeoOfficeJ) on my Mac, as well as Microsoft Office. Sometimes, one does the job better than the other. (For example, I use a lot of movie clips in my presentations - I need something to make me a little less boring of a presenter - and OpenOffice still doesn't support embedding video. At least, not without a trained expert to guide me, which I don't have.)
Anyway, I just clicked on Geoffrey Moore's presentation in my mail reader, and NeoOfficeJ decided to open it. I never told it to, and I never set it as the default reader of .ppt files. I like to choose the best program for a given job, and not have the programs try to automate choices like this for me. I'm pretty annoyed at the NeoOffice team over this....
My advice? Make a great program, and let people choose to use it. Don't force your way into their usage routine.
I just received the slides for Geoffrey Moore's OSBC keynote. It looks great. Then again, I tend to like anything that applies philosophy/history to day-to-day problems. In this case, Geoff is going to apply Darwinian theory to open source's "evolution," and draws some lessons from Darwinian natural selection to guide strategic decisions by IT vendors and buyers.
If you haven't yet registered, you should. It's going to be a high-octane event.
David Berlind has a great interview with the ever-quotable Michael Tiemann of Red Hat. (At an HBS conference where I heard him speak on a panel, he started off by declaring that we live in a fascist state. I can't remember why, and my life hasn't seemed all that fascist in nature since then, but I'll give him the benefit of the doubt.) In case you missed it, Tiemann is the president pro temp of OSI. Tiemann is also VP of Open Source Affairs at Red Hat (where he does a great job - ever heard him speak? He's great).
The question is, however, whether he can serve both masters equally well. He's a fantastic partisan for Red Hat (which I mean in the best sense). How well this will translate into being an impartial advocate for open source through OSI is a question that begs to be answered. It wouldn't normally be a big deal, except that Tiemann is such an ardent (and successful) partisan for Red Hat, first, and open source, second. (My opinion.)
Danese Cooper and others walk this line easily, because their primary interest seems to be building the open source market (and expecting their companies to gain significant shares of this growing market). I put myself in this camp, btw - I manage the content for the Open Source Business Conference, but my Novell affiliation in no way affects who speaks, etc. In fact, Novell probably gets negative treatment from me because I don't want my employment to affect the conference.
Red Hat and its employees, however, sometimes strike me as having this community impulse somewhat muted. Not Havoc and many of the engineers, but rather others in management. Of course, this is exactly what they should be doing: their duty is to their shareholders, first. (It's the law, as Henry Ford found out when the Dodge brothers sued him a half-century ago for wanting to lower profits to benefit customers.)
But that duty ends where OSI begins. The minute Tiemann puts on the OSI hat, his role for OSI is to maximize its effectiveness as an organization, whatever that may mean for Red Hat, or step down. Tiemann may surprise me on this. I hope so. I'll happily eat my words (digitized though they may be). Either way, I wish him the best of luck.
A maelstrom of discontent has recently arisen over open source license proliferation. Martin Fink kicked it off with his Linuxworld keynote (available here, wherein he said:
Many people don’t realize that there are dozens and dozens of open source licenses today. There is really no value, and much confusion in having that many licenses. If you’re a vendor and you plan to create a new license: Stop, please don’t.The argument is that open source license proliferation breeds uncertainty (license incompatibilities, as well as simply knowing what the various licenses say/mean) which, in turn, creates costs and slows open source software adoption. Don Rosenberg, author of Open Source: The Unauthorized White Papers has written persuasively on the problem, which I blogged here.
This is very random, but I've always associated Larry Lessig (my old advisor at SLS with Billy Corrigan, former lead singer of The Smashing Pumpkins. And Moby. I have no clue why two bald musicians remind me of Larry, but there you are....
"Despite all my rage I am still just...
going to the west side
weapons in hand as we go for a ride...
looking for the Supremes and copyright bigots?"
I probably shouldn't reveal the strange associations I make.
Dave Rosenberg (used to run Comdex, and now does Linuxworld) has a great post on how lame most conference speakers are. I talked about this phenomenon a bit earlier, from a different angle, in one of my posts. It only takes attending a few conferences before you realize that either most speakers have zero to say, or they're so heavily media trained (as Dave points out) that they aren't allowed to say anything more interesting than their names. And most of those are boring, too.
No one wants sales pitches. No one wants to endure a product data sheet masquerading as a keynote. It doesn't sell licenses.
Maybe that's why I don't go to media training at my company. Or why I studiously avoid telling the marcomm people when I'll be speaking. I want to be able to say interesting things, or at least make an attempt at doing so. The higher in an org you go, however, often the less interesting the speakers become. Maybe it's because the higher up you go, the more managerial (and less entrepreneurial) people become. Or maybe it really is that executives forget how to think and speak informatively, being so heavily primed to stick to the company line (even if that line is a loser).
