March 2007

Okay, I’ve spent way too much time playing around with the visualizations on this IBM alphaWorks site called Many Eyes. I read about it on CNet, and popped over to take a quick look…over an hour ago!

I know I’ve seen at least one user generated “data reference” site where users upload charts from various sources. I can’t find that site again for the life of me though.

The interesting thing about Many Eyes is the visualizations you’re able to create with the data, and how others can create different visualizations with the same data.

This is a cool one dealing with US Budget Expenses from 1962-2004. Check it out!

Roberto writes that Sourceforge (SF) is adding the ability for OSS projects hosted on to sell support & services, and for users to buy these services. SF plans to launch Marketplace in late spring/early summer.

Kudos to Roberto for not only blogging about the Marketplace announcement, but also for finding the Senior Market Management position for SF Marketplace and linking to VA Software’s (SF’s parent company) latest 10Q. It’s amazing that they made $13M in e-commerce revenue in the most recent quarter from selling consumer goods through ThinkGeek. Also interesting is that VA made $3.8M from online media, mostly advertising. I would have guessed sites like and would drive more than ~$4M in Ad revenue per quarter. But maybe geeks don’t click banner ads because they use Firefox with Adblock Plus ;-)

Anywho, so now it gets interesting between SF Marketplace & Red Hat Exchange, although they appear to be going after different projects & customers.

Red Hat is definitely going after the enterprise customer who has (by in large) made the decision to run the given product on RHEL. So, the testing and certification of a product with RHEL is critical to being included in the Red Hat Exchange.

But a majority of customers don’t run OSS on RHEL. A lot of folks run on Windows. Even JBoss use on Windows was 50% or so, which is somewhat lower than similar figures we saw from our WAS CE downloads about a year ago – things may have changed since though.

Today, SF is a leading destination for finding & downloading OSS products. The SF Marketplace appears to be the next logical step.

For the large projects on SF who already have support & services business attached, I doubt that this announcement changes much at all. Customers interested in, for example, Hibernate support, will find their way to and then to So I’m not sure why RH/JBoss would want to give up a percentage of their revenue to VA/SF just to get included in the SF Marketplace.

For emerging projects or for projects with a small development team/community, a majority of the 144,548 projects on SF (i.e. Longtail projects), getting included in the Marketplace would make a lot of sense. Why spend time and resources building & managing the support infrastructure and marketing your support & services if SF can provide that to you for a (small?) cut of your revenue.

Experience tells me that customers are cautious when it comes to spending money. When they do, they want to spend with vendors that have a strong future. So, for longtail projects on SF, I’m not sure that the SF Marketplace will change much of this customer behaviour. Remember, downloads are a badge of credibility, something that longtail projects lack as you move further down the tail. So, if the SF Marketplace won’t drive a lot of revenue for longtail projects and the large projects on SF already have a Support & Service offering through the vendor site, what’s left for SF Marketplace?

I could be proven wrong, and maybe users of longtail SF projects will decide to fund future development through support purchases.

I wonder if SF Marketplace will allow me to see who is offering support & services for, say, Hibernate in one place (i.e. JBoss, Virtuas, Neward & Associates, etc.). But maybe the project owner would have a say in who can offer support for said project? :-)

It will definitely be interesting to see how customers adopt SF Marketplace & Red Hat Exchange.

I haven’t posted in a little while due to travel (who says offline RSS readers aren’t necessary) and my attention being elsewhere. But I’ve been catching up on some reading.

First I read this very interesting post from Shaun Connolly from JBoss in response to my questions about JBoss community.

Shaun says:

“Open source communities extend beyond those who interact directly on the open source projects, mailing lists and forums, and include the users, customers and partners in a wide variety of ways. In my opinion, there are neighborhoods within the larger community that have their own perspectives and ways of interacting with the larger community.”

I really like the neighborhood approach that Shaun suggests – I’ll need to think more about this one. I’m just not sure how or why third party users, customers and partners would be differently motivated than third parties who “interact directly on the open source projects”. Aren’t third party contributors users or partners, or don’t they sit inside of a customer shop?

The point that Shaun makes about how Red Hat serves expands and recognizes their community could very well have been written by IBM, Oracle, Microsoft or any other traditional software vendor. Not surprising considering how Red Hat tracks with traditional software vendors in many aspects of their business.

Then I read “My thoughts on communities backed by companies” from Jordi Mas, (linked via Matt’s Infoworld blog). Jordi poses a few questions to rate how community driven your development is (regardless of whether you’re open source or not, because, as Jordi points out, Microsoft and other traditional vendors also have strong user communities).

Jordi poses questions like

“Is the product roadmap publicly available?” or “Do developers from outside your company have privileges to make changes on the code base based on their knowledge? (Meritocracy) Or are they second class citizens?”.

I like the questions, but what’s missing (and likely always will be) is a way to assign an importance rating for each question. So, what if the product roadmap isn’t publicly available, but commit privileges are dolled out based on merit? Does that make the company in question less, the same, or more “community driven” than a company with a public roadmap that only gives commit access to employees?

But clearly there is a difference between a community found at Apache, Eclipse or than other communities. It’s just not easy to define.  But, that doesn’t mean we shouldn’t try.

After speaking with many vendors on the AjaxWorld expo floor about their business models it seems like the OSS business model is being used by about 50%, for the products here at AjaxWorld.

Selling applications or custom application development services:

Product is not OSS or does not use OSS business model in Web 2.0/AJAX/RIA space
Isomorphic Software

OSS product or uses OSS business model in Web 2.0/AJAX/RIA space

I was somewhat surprised with the high share of companies using a traditional software license business model. Of the companies using the OSS business model, the best known, Laszlo, is totally focused on selling support around their open source framework/foundation, they’re selling end user applications. They built a framework to simplify writing applications and made the strategic choice to open source the framework to deliver value to 3rd parties also. The openLaszlo framework drives services opportunities (and later, customized applications), but very few customers purchase support for the framework itself.

I was expecting more OSS companies (although ~50% is pretty high). It could be that these companies got into Web 2.0/RIA/AJAX/etc. 2+ years ago when OSS adoption was less than it is today. It could be that there’s room for both models?

I purposely didn’t write about the Exadel partnership when it was announced as I wanted to get some more details on it. A week has passed and little has been written that furthers my understanding. And I must apologize for not emailing someone at JBoss to get some clarification (it’s been busy here). As such, I ask you to read the comments before finishing this post, in case one of my JBoss readers writes in to correct my take on the deal.

Here is the quote from the TSS discussion on this deal:

Sacha Labourey: “The Red Hat and Exadel partnership is twofold. First, Exadel will open source its commercial products — Exadel Studio Pro and RichFaces — at as Red Hat Studio Developer and JBoss RichFaces, respectively. Exadel is also moving its popular Ajax4jsf project, currently hosted on, to, where it will become JBoss Ajax4jsf. Second, Red Hat and Exadel will continue developing the three projects going forward, including integration with existing JBoss platform technologies such as JBoss Seam. This ongoing development will be done under JBoss’s leadership.”

Then, JBoss/RH’s Gavin King comments:

“….. Seems like their (i.e. Exadel – added for clarity by Savio) business has actually worked out pretty damn well, given that we just paid them a whole bundle of money to get this deal done. ;-) This is not a typical case of a commercial vendor deciding to give away a product they couldn’t sell, in a last-ditch attempt to get users. Rather, this is an acquisition of successful technology, for money. It follows the model of JBoss Transactions, which was acquired from Arjuna, and JBoss ESB, which was acquired from Aviva. …..”

Am I off in left field when I think that this is little more than JBoss/RH acquiring technology from Exadel for a price? That the “….decided to open source the products at” parts of the announcement are more about the cash that changed hands and less about validation of any community site? Reading Gavin’s comments leads me to think as much.

Don’t get me wrong, there is absolutely nothing wrong with that, acquisitions are a way of life in the software industry.

What doesn’t sit well with me is that the TSS Q&A with Sacha and the JBoss/RH Q&A about the deal don’t really clarify if:

  • (A) Exadel decided to open source their products and choose the JBoss/RH community because of its strategic fit with Exadel goals or for whatever reason that Exadel would feel that one ‘community’ is better than another ‘community’
  • (B) Exadel was paid to transfer their copyrights to JBoss/RH, and the open sourcing of Exadel products at JBoss/RH was part of the deal

I’m not the expert, but I don’t think that the Apache Software Foundation or Eclipse ever paid a 3rd party to decide to open source a product into the Apache or Eclipse community. When IBM acquired Gluecode, (which was building products on top of Apache Geronimo), we positioned the deal as what it was, an acquisition. Nice and clear.

Can anyone at JBoss/RH shed some light on who owns the copyrights to the code going forward? Because that should be a simple way of knowing if this was a 3rd party endorsing the JBoss/RH community or a 3rd party selling their IP to JBoss/RH.

Either way, it’s a business decision. But I don’t think it’s cool to paint an acquisition of technology as a partnership that endorses a community/external development site.

I’ve been catching up on reading and found this on Roberto’s blog which pointed to this story on Matt’s Infoworld blog.

One interesting point that Roberto highlighted was Matt’s comment:

“They would buy from us, anyway, even if we gave them the code. Not all of them, mind you: most of Europe seemed to be running Alfresco (Community – then our only open source product) without paying us a centime. But we figured that these companies wouldn’t buy from us, anyway.”

Well put Matt. What’s wrong with a set of customers who don’t pay for using your open source software? Non-paying customers don’t directly cost you anything other than *maybe* a few dollars in bandwidth.

But they provide significant value.

Non-paying customers are users first. So, they may find and report bugs (although these users likely won’t get an expedited response) that would also impact your paying customers. So they help with product quality.

From a competitive standpoint, isn’t it better for a customer to be using your open source product without paying, than using your competitor’s product?

Depending on the openness of your community and the development practices at the non-paying company and their industry, you may receive some fixes or new function from this customer set.

But most of all, these non-paying customers act as a credibility badge for your product. How many of us search for some tool, say a password keeper, and finding 10 of them, narrow our review process to the top 2 based on the number of downloads? Try as our parents did to teach us otherwise, “peer pressure” still impacts a large portion of everyday decisions.

Which brings me to this HBS Working Knowledge article titled “How Do You Value a Free Customer?” The research behind the article is more focused on multi-sided markets, i.e. companies that serve two or more distinct customers who need each other to form a market, although the company may only drive the majority, or all, of its revenue from one party. The author provides EBay,, or the Real Estate market as examples. Traditional models of CLV (customer lifetime value) don’t take into account the value of, say a buyer, if the seller is who pays. As the author puts it, traditional models ignore the network effects.

Vendors (OSS & traditional) focused on converting non-paying customers shouldn’t lose sight of the value these customer provide.

Next Page »