Linux Today: Linux News On Internet Time.

Community: Purpose and Participation: It's Harder Than It Looks

Aug 31, 2004, 00:15 (0 Talkback[s])
(Other stories by Gary Edwards)

[ Thanks to Gary Edwards for this article. ]

Recently OpenOffice.org member Roger van Vissingen responded to a discussion thread entitled, "Microsoft fleeing up market; encounters networked market inefficients." Unexpectedly he broke from the "Microsoft is the center of the universe, but not for long" mold of commentary, and posted the following quote:

<Roger> ... What is not yet clear is what are going to be the successful open source ecosystems. Peter Drucker said the purpose of business is to create a customer. The mistake of mature businesses is to try to MAINTAIN their past glories past irrelevance, and fail to create new customers for whom they do have relevance...<.../Roger>
full thread here ...

Hummm. "Successful Open Source Ecosystems." "The Purpose of Business." Good topics. And Roger managed to fit into the same paragraph a comment from Peter Drucker, last centuries Über corporate management guru. While I've done enough Microsoft commentary (i.e. "bashing") to last four lifetimes, this new angle is interesting. Especially since every corporation outside of Redmond seeks an insight into how to make a go of participating with a successful open source ecosystem. Open source is a competitive advantage Microsoft can never touch. It's like doing business in waters that Redmond can't navigate or mine. If only one knows how to use it.

There is an interesting "must read" post from Simon Phipps (Sun) concerning open source ecosystems, and how to participate. It's titled, "The Subscription Model: A Necessary Trend for Open Source Deployers."

Based on discussions that took place at a rather informal JavaONE meeting I was fortunate enough to attend, Simon artfully argues Sun's approach to participating in a "successful" open source ecosystem.

1. A New Organizational Model

Roger's "Drucker quote" is spot on, but I think we could benefit by expanding the Drucker philosophy a bit in hopes of meshing in with Simon's effort to throttle the pulse. Peter Drucker has also said that the corporation is the premier model for organizing people, resources, and finances for the purposes of profit. In this context one might argue that Open Source Communities are a new organizational model, the purposes of which are services and contributions being shared instead of sold for profit.

Another interesting aspect is that these new organizational models can fully leverage the enormous impact of Metcalf's Law, "the Network Effect." In fact, it's hard to determine which coalesced into a driving force first, the Open Source "shared" contribution model, or the "Network Effect." Could one exist without the other? Not that an answer to that question matters. What really matters is the clear observation that Drucker Corps style vendors can't tap into the "Network Effect" supernova without having to seriously compromise their core organizational principles. For anyone still pursuing Gatesian style monopolist perversions of the Drucker model, this realization is disastrous.

Simon outlines some of the reasons FLOSS will rule the day, but his focus is clearly on how it is that a Drucker Corp can "profitably" contribute to and participate with OSS orgs. The conflicts of organizational purpose looms large. Interestingly these reasons seem to center on advantages important to computational consumers, or "customers"--or in the world of FLOSS--the members (or beneficiaries) of OSS Communities that Simon refers to as "deployers." Simon also explains how corporations can and do participate with FOSS to become important delivery channels for these advantages.

<from Simon> "What freedoms do deployers need? They're actually rather traditional:

  • Function that meets actual business needs

  • Freedom to change supplier, so that prices can always be negotiated

  • Freedom to choose new software solutions as business needs evolve

  • Control of the data, by implication control of the format it's stored in

  • Protection from liabilities associated with the development of the software" <.../Simon>

2. Ownership & Independence

In many OOo discussions these advantages are often referred to as "independence" or "ownership" issues. OpenOffice.org and Mozilla.org are about as fine an example of what is meant by first class organizational participation in a "successful open source ecosystem" as can be found. They've mastered the demanding architectures of participation and interoperability. So much so that vendors the likes of Novell, IBM, and Sun have been able to extend that interoperability into highly integrated product lines. No doubt there is great value to these "professionally edited" integrated enhancements. And one can see the Network Effect being harnessed.

So far so good. But there's a catch. Call it the curse of the Network Effect. Remember the Law of Drucker, " the purpose of business is to create a customer?" When you dance with OSS communities, what you create is a "member" not a customer. Ooops.

Maybe a vendor profits from these "members." But unless the vendor can master the art of what former BEA CEO Bill Coleman once called "running the stack", there's a rising cost to repurchasing these members. "Running the stack" is a process of open innovation at the higher levels of systems enhancements, while pushing past innovations into the lower foundation of open source--open standards layers. Once clients trust you to "openly" innovate, they'll subscribe and upgrade as you run the stack with enhancements targeting higher levels. Just don't compromise their ownership and independence (or what Jonathan Schwartz calls the empowerment of "substitution"). You'll never see them again. You'll never get another bite of that apple precisely because "membership" has Network Effect based "ownership" advantages.

From a global technologists perspective we frequently hear statements something like this: In a distributed network centric world, loosely coupled systems lead to greater independence at all layers of connectivity, including the end points. Independence of platform, application, and proprietary middleware, leads to consumer/client control (ownership) of both the information, and the information process. Woe to proprietary vendors trying to insert control points, erect barriers, and self righteously compromise the Internet formula of open standards, open interfaces, open communications and messaging protocols, and now, open XML technologies.

IMHO, we can get beyond the goop of the 30,000 foot view and drill down to some clear "ownership" specifics simply by looking at three architectural aspects of the OpenOffice.org effort: The OASIS OO XML file format, the UNO Component model, and, the OOo architecture of participation (to borrow a phrase from Tim O'Reilly). (Apologies also to that other great champion of cross platform magnificence, Mozilla.org, in that my focus here will be on OOo)

3. Ownership at the File Format Level

Because the OASIS Open Office XML file format separates the information content and context from application and application platform, OOo members fully own their information at that lowest of all levels, the file. They don't need any vendors permission or licensed assistance to take that information with them when they switch applications or platforms. Nor do they need vendor permission to collaboratively work that information across other applications and other platforms. Because the OASIS OO XML file format is both human and machine readable (structured), and fully compliant with the growing ecoverse of open XML tools and technologies, universal interoperability with applications of all sorts becomes a trivial matter. A statement that will be just as true 200 years from now as it is today.

These extremes of universal portability and endless interoperability leave us with a glowing conclusion, the likes of which OOo marketing is still trying to come to grips with and explain: the OASIS OO file format redefines "ownership" for the next generation of collaborative computing.

Trying to explain this to someone who can't imagine the possibilities of an independent life outside the bounds of the Microsoft box is difficult. Trapped by their legacy investment in the platform, the hopeful tendency is to see the future through the dark glass of whatever Windows can do today. Or worse, through whatever Microsoft promises Windows will do tomorrow. One thing that helps though is to explain the importance of XML in terms of structured vs. unstructured information. The trick is to get outside concerns about the costs of leaving the Windows platform and focus on the ownership costs of remaining there.

This reel has worked for me: "It's estimated that over 90% of mankind's knowledge is in an unstructured format. That means that the information is human readable, but not machine readable. Which in turn means that since the information is unstructured, we can't apply the advancing power of our "knowledge machines", our computers. In the information age that's the equivalent of trying to dig the Panama Canal with a pick and shovel instead of using nitro, mammoth bulldozers, cranes, advanced hydrolic systems, and giant trucks. OpenOffice.org is the only office productivity suite available today that natively generates structured XML files. And OOo can do this for the most complex "compound documents" imagined. The Australian Historical Society is already converting their knowledge storage to OASIS OO XML file format so that these massive volumes will be computationally useful to future generations."

My point is that if you speak to the issue of the Windows box, the MS platform and MS applications, your's is likely to be just a voice at the end of a long, very dark tunnel. If you speak to "their" information though, even the echo is deafening.

This level of information independence / ownership also means that members can collaborate without having to overcome the barriers of expense and hassle so endemic of permission based ecosystems.

The Universal Component Model:

The open XML file format is only part of the ownership story. OOo members not only take ownership of their information, they take ownership of the "information process." The other parts are provided by the open UNO component model, and, the responsible attitude of the OOo Community.

The open UNO component model carves the office suite into easy to grab components that developers can access, re purpose and use to create new services and inter application connectivity that might go far beyond anything the OOo engineers and designers ever imagined. Indeed, IBM's next generation WorkPlace configuration does exactly this sort of OOo component based re engineering.

The funny thing is that with UNO, IBM is able to achieve measures of interoperability and integration between WorkPlace and Lotus Notes, the likes of which they were unable to achieve with the Lotus Office Suite! Notes defined collaborative software for the client/server platform model. One has to wonder what will happen if IBM unleashes a web based collaborative workflow model using the OOo WorkPlace compound documents on one side, meshing with Notes "intelligent" workgroup documents on the other? Very cool.

WorkPlace exposes some basic truths about what it takes to achieve extremes of interoperability. First, OOo is an extremely flexible XML engine. Second, information processing with OOo components is not application bound, but able to integrate more fully into the much larger context of the members complete productivity and business processing environment. What looks to be web magic is really just a well thought out next generation component architecture flexing her muscles.

Microsoft applications have always held this "integration" advantage by virtue of the facts that MS controls the OS, the Win32 API, and every other application users might need integrated into a customized productivity environment. The sad truth is that in spite of the DOJ's anti trust settlement, Microsoft continues to own, by permission or license, the desktop productivity environment. And because of the settlement, MS is legally now able to extend that ownership into desktop connectivity with servers and devices. Even sadder, the settlement effectively eliminates any possibility of a for profit corporation competing with Microsoft on the Windows platform. Baring a giant leap to Linux, productivity alternatives will come from OSS communities.

While IBM, Novell, and Sun are relentlessly pushing OOo components and services into highly integrated "productivity environment" alternatives, I think it's great news that the core OSS communities are finally getting together to start exploring the perfection of their own integrated profiles. While it seems every OSS community is driven to pursue extremes of interoperability, I think there also comes a time when it makes sense for core communities to coordinate efforts towards perfecting their own integrated solutions. Why should we leave cross community integration opportunities to corporations and distros? Besides, setting an integration threshold might serve to push our corporate benefactors further up the stack, and beyond the limited expectations of providing alternatives to MS.

There is another important element of passing ownership of the information process to members. In his blog Jonathan Schwartz discusses the issue of "substitution." A vendor might provide an integrated stack based on open standards and comprised of mostly open source solutions. But that doesn't guarantee that the components can be swapped out and replaced by alternatives without the also collapsing stack dependencies, compromising interfaces, and breaking critical protocols and methods. It's not "substitution" if everything else breaks.

One might also say that "substitution" is directly correlated with good OSS citizenship. Sun is of course both the OOo benefactor and driving force. Novell has returned all of their enhancements, changes, and interoperability extensions to OOo. We're still waiting to see how IBM plays it. No wonder Sun chose the GPL for their latest OSS contribution, "Looking Glass."

The Architecture of Participation:

Important to the "ownership of the information process" discussion is this element of how a community goes about taking responsibility for the future. This is more than the licensing issues Simon points to. It's also more than the good citizen / good stewardship issues Jonathan Schwartz references. IMHO, "responsibility" has to be both built into the community framework of participation, and, become a torch that simply refuses to die. For members to truly take ownership of their information process, they need to believe in the strength of the steel that binds a community to triumph in that race to the future.

Fortunately our benefactor Sun was able to bake into the OOo framework the magic of universal participation. It took Mozilla.org many revs to finally strike that balance of a component based framework where similar levels of collaborative participation and interaction between end users and the masses of interested developers began to occur. Now the core communities are finally at that point where they can make that important shift from "application" focus, to focus on an integrated productivity "environment." One that developers can reliably target with solutions that mesh without barrier into the information processing systems. Meshing without compromising user ownership. The recent announcement by OOo and Mozilla communities that they would cooperate on perfecting an integrated solution is the first sign that we are entering a new era, the era of the portable rich client environment. The era of interoperability above the distro.

The Shadow of a Monopolist:

I was fortunate to attend the informal JavaONE meeting hosted by Simon and Tim Bray, and referenced in Simon's article. Curtis Sasaki (Vice President of Engineering, Desktop Solutions Sun Microsystems, Inc), and Peder Ulander (Director of Marketing, Desktop Solutions Group Sun Microsystems, Inc) presented the Java Desktop System. I hope Simon writes about other aspects of the discussion that took place. The issue of Sun providing a "portable JDS" environment is very important to anyone wrestling with the myriad difficulties of devising workable migration strategies. Strategies that start with an ownership empowering slide into a OOo/Moz/SleepyCat style productivity environment (JDS) that can also interoperate and transparently connect with Linux Server based alternatives. From there it's easy for a solution provider to recommend that new computer purchases be based on JDS. Knowledge workers will hardly know the difference. Businesses will save money. Ownership and independence will be achieved through a graceful transition away from vested legacy systems critical to ongoing business processes.

Sun desired to present the JDS as a solid desktop system, sold at the right price, backed by credible service and support. IIRC, Simon's "editing" comments were made in response to two blanketing suspicions. The first was that of the JDS as an alternative to Microsoft. Not the Windows desktop, but an alternative to Microsoft the juggernaut. The second was the basic "how open is open" concerns about the JDS. Sun itself has raised this concern by making comparisons to Red Hat. Accusing Red Hat of proprietary lock-in, of devising methods that lock applications into their proprietary configurations of Linux.

Simon's "editing" comments were originally intended to deflect the accusation that Sun would similarly use the JDS as a proprietary stalking horse. A Gatesian trap waiting to be sprung on hapless developers and easily blinkered computational consumers. In addressing these suspicions, Simon argued that all Sun had done was to carefully "edit" existing open source components, selecting the best of breed, trimming the graphical interface, and introducing simplified systems management features that in no way compromised the stacks high substitution ratios that customers value so much.

Sun's record with OSS is that of an honest broker. While their not 100% open source in everything they do, where they have engaged open source communities it has been with honest intentions and openly transparent contributions. Incredible contributions! And they haven't deceived or ticked anyone. Yet, they can't do anything without first answering to persistent suspicions. Maybe that's a good thing. The price of freedom being eternal vigilance. I only wish the standard were equally applied to everyone.

One thing's for sure. Microsoft has truly poisoned the water. It's near impossible to present a highly engineered system without someone (everyone) accusing you of secretly setting a Gatesian trap. Nobody wants to get caught in another trap where the cost, effectiveness, and capabilities of their information systems are hostage to the arbitrary permissions, limited provisioning, and restrictive licensing methods of a single vendor.

Ownership. It's a good thing.

Gary Edwards is an OpenOffice.org volunteer serving on the OASIS Open Office XML TC

Related Stories: