Panel discussion on Open Source business models

freelock writes "The MIT Enterprise Forum of the Northwest is hosting a panel discussion about Open Source Business models this coming Wednesday, February 16, in Bellevue, WA. The panel will be moderated by Dan Woods, author of the upcoming O'Reilly book Open Source in the Enterprise. Panelists include: Jacob Taylor, co-founder of SugarCRM; Byron Sebastian, founder and CEO of Source Labs; Kevin Foreman, general manager of Real Network's Helix project; and Tom Brubaker, the VP of a security company and founder of a demised open source company. Should be a very interesting discussion! Disclaimer: my company, Freelock Computing, is sponsoring this event.
Sign up here!"

Comments

Sveasoft

[Bios_Hakr]

Want a good buisness model? Take someone's GPL code. Modify it a bit. Begin charging to access the "perpetual beta" code. Promise to let people have access to the source once it's finalized...only it'll never be finalized.

To top it all off, threaten, DMCA, and generally shit all over the GPL and it's guarentee to allow everyone to redistribute.

Last I heard, they are rewriting everything as NVIDIA style modules to augment the GPLed firmware.

Anyway, it's really pissing in the pool of anyone who has ever contributed to the Linux kernel...and it's all technically within the GPL.

Re:Sveasoft

[Jack Tanner]

Want a good buisness model? Take someone's GPL code. Modify it a bit. Begin charging to access the "perpetual beta" code. Promise to let people have access to the source once it's finalized...only it'll never be finalized.

I've been watching the Sveasoft saga from the sidelines, and I must say: I love the fact that a company based on a cool hack, a lot of sweat, and a real need for its services is doing well, and is doing well with the help of Linux. At the same time, they're going about it in an insanely antagonistic way. I'm surprised Slashdot hasn't been served with a DMCA notice for hosting TheIndividual's Journal. (Or maybe it has. How would we know?)

Sveasoft could be so much more positive! MandrakeSoft actually does follow the "release to members first" model, but because it holds to its promise of "it'll only be for club members for a very little while", Mandrake is at the same time popular, appreciated, and highly karmic.

Sveasoft, on the other hand, has a version of its Alchemy firmware -- 6.0rc6 is the latest -- that it calls beta, and therefore not open to the public. At the same time, it's started development and is promising features from Talisman, the next release, which is obviously not open to the public either. Things would be so much simpler if, say, Sveasoft actually announced in advance when a release would become publicly available.

Apple Open Source Business Model?

[2nd Post!]

Is anyone going to comment on the Apple business model?

Open source core for OS X (aka Darwin) but closed GUI and secret sauce?

Or open source for their Darwin Streaming Server?

Open source for ZeroConf/Rendezvous?

Open source usage of CUPS? Of X11?

Using the open source KHTML rendering library with a closed source browser, Safari?

I'm sure there is more.

If anyone wants to 'discuss' any examples, please provide linkage proof of violations or usage. Here is a link to Apple's open source projects pages.

Re:Apple Open Source Business Model?

[Surye]

Using the open source KHTML rendering library with a closed source browser, Safari?

I've always wondered how they managed this without being in breach of the GPL... I mean, if it's linked against it.. Unless of course it is LGPL, or not really linked.

Re:Apple Open Source Business Model?

[2nd Post!]

LGPL is mentioned in the .h files.

I want video.

[Jack Tanner]

Will there be a video stream of the panel? Or a downloadable recording post-factum? Or at least a transcript? Inquiring minds want to know.

Open Source vs. Gillette

[G4from128k]

It seems to me that the core of the open source business model is not that different from Gillette's razor-blade model. There are three parallels. First, in both cases, an expensive asset is provided at less than cost (software code or razor handle). Second, the same people that create the asset also offer some service or consumable (software customization/support or razor blades). Third, the enterprise then thrives on revenue from supporting the use of the freely provided asset

Admittedly, there are large differences between OSS and Gillette. The first set of differences are interrelated and driven by the cost of delivering a copy of the asset to the customer. Gillette faces a high cost of creating razor handles. Thus, it must create captive customers who are forced to buy blades. With OSS, the low cst of copies (the cost of download) means that OSS can tolerate a very high percentage of leaches (customers that take the free stuff and never pay for support or contribute). The ability to tolerate of leaches is a key prerequisite for Open Source.

Another similarity between OSS and Gillette is that both create (to varying degrees) an ecosystem that allows for enhancements to the core asset (e,g., add-on utilities for OSS or additional cobranded products for shaving). The fact that the core asset creates a hub for value attracts additional resources and additional customers in a virtuous circle.

Of course one key difference between OSS and Gillette is in the creation of the original core asset. Gillette pays a very large sum up-front to design the system, build the factories, and create the product (Gillette spent $1 billion to create the Mach III and 2.5 billion to create the Sensor). OSS pays "nothing" to its developers in the pre-launch phase.

The reason this works for OSS is two-fold. First, the cost of the core asset (the codebase) is almost entirely in labor. The ubiquity and low-cost of computers means that anyone can become a developer. This works because people seem psychologicaly more likely to donate time rather than money (a developer might spend 20 hours week on an OSS project at nights and on weekends, but would never consider donating 50% of a years salary to the project). Second, the very same people who donate their time to create the codebase have some hope of reward from the follow-on processes (customization and support) -- for Gillette the person the makes the handle is different from the person the make the blade.

Based on this, I would say the Open Source applies to endevours that have the following characteristics:

Low cost of copy of some core asset

Some potential for process-related revenues for use of the asset (customization, support, subcriptions, consumables, etc.)

Labor-dominated cost of asset creation

Interchangeable labor between the asset-creation and consumable/support sales process

Re:Open Source vs. Gillette

[Anonymous Coward]

No one company can control the sale of blades though, so the analogy isn't completely apt

Re:Open Source vs. Gillette

[Ed Avis]

First, in both cases, an expensive asset is provided at less than cost (software code or razor handle).

Nope, a Linux CD sells for about a dollar, and that's about what it costs to manufacture.

Re:Open Source vs. Gillette

[G4from128k]

First, in both cases, an expensive asset is provided at less than cost (software code or razor handle).

Nope, a Linux CD sells for about a dollar, and that's about what it costs to manufacture.

And, if a person downloads Linux of someother OSS, how much do they pay? Bandwidth may be cheap, but it is not free.

I'm sorry my post was not clear, but I was speaking of more than just the cost-to-copy. I was speaking of the total cost to provide the code. The full cost to provide a million copies of Linux (if a private company were to attempt to recreate Linux all by itself) is much greater than the cost of a million blank CDs. It includes the effective cost of all the "free" labor contributed by OSS coders (the resources required to create that first copy). If Open Source can be a business model, it must find some way to reward or motivate those who contribute their time to the original creation of the open source good.

The point is that full cost to provide something includes both the cost-to-copy (variable costs) and the cost to create that first copy or to create/maintain the infrastructure (fixed costs). If an enterprise cannot find a way to amortize its fixed costs across its activities, it will go bankrupt.