Freedom vs. Free

Open source software is definitely not free when you look at the TCO. I have said it and will repeat it: I like to look at open source from a business perspective and believe much more in Marc Fleury's Professional open source that Richard Stallman's. Just like communism, if everything about open source is free, it will ultimately collapse. The reality is engineers need to feed their families. I have nothing against companies (like JBoss or Orbeon) charging for support and maintenance services (the scalable and predictable portion of the software business).

The good news is that the whole industry today understands open source much better; and when asked what's the most important advantage in using open source, people are not jumping to the old answer: it's free! Or, it's cheap! Many studies show statistics illustrating this shift, the latest I found was conducted by Computer Economics, here are their conclusions. The gist of it is that only 22% thought that it was cheaper to adopt open source while 44% chose "less dependence on vendors". In other words, people are interested in independence and freedom more than free or cheap. I also found a definition of Free Software on GNU and it had nothing to do with price, it talks about four kinds of freedom:

- The freedom to run the program, for any purpose
- The freedom to study how the program works, and adapt it to your needs
- The freedom to redistribute copies
- The freedom to improve the program, and release your improvements to the public, so that the whole community benefits

What's important to you (or the organization you work for) when considering open source solutions?

Michael Dell likes Red Hat

At age 40, Michael Dell's personal wealth is estimated at $16B. Dell gave his initials to a NYC-based investment company MSD Capital which manages $10B of his fortune by investing in real estate among other areas. This week, news emerged that in January of 2004, MSD Capital decided to spend $100M to buy 3.9 million RHAT shares. Remember, Dell Inc. resells Linux software on its servers and was an early investor in Red Hat (a direct competitor to Microsoft). At first, given my respect for Michael Dell as a visionary and a business man, I read this transaction as: Red Hat is onto something very interesting. But when you think about it, MSD Capital may have made the decision to invest without Dell's involvement or blessing.

Another question I had was: does this mean that Dell Inc. is preparing the acquisition of Red Hat? But they have never acquired a software company. Would it be a smart move? I don't think so, the majority of PCs sold by Dell still run Windows, the last thing Dell wants to do is upset the Redmond titan. Unless Microsoft buys Red Hat, wouldn't that be interesting?

Also, MSD made other investments in completely different areas such as restaurant chains IHOP and Steak n Shake. So Red Hat could simply be part of a diversification strategy. Finally, since MSD's investment, RHAT shares dropped by about 45% percent. So may be there is not much to conclude from MSD's $100M move.

Read more about the Dell Inc. / Red Hat corporate alliance.

From the horse's mouth

Two days ago (5/10/05) I wrote a piece on IBM's acquisition of Gluecode, an open-source play that they are planning to use to open low end markets before converting Gluecode/Geronimo users to IBM's enterprise application server WebSphere.

One of the questions I was hoping to get answers for was: did IBM feel too much heat in the SMB market from JBoss and decided to gobble up a start up to hurt JBoss by making Geronimo better and more popular?

I guess I was too optimistic thinking that I had so many readers that all my questions were going to be answered and that's the end of it. Sadly, this morning my question was still unanswered so I decided to look around for possible hints. Who better than Marc Fleury (who loves to talk about JBoss's success and enjoys commenting on his competitors' every move) to answer my JBoss related question? Let's get it from the horse's big mouth!

Marc had already used his blog to share his initial reaction to IBM's acquisition of Gluecode. In addition to that, I found an audio interview of JBoss's CEO by ZDNet's David Berlind; it is available in MP3 format if you are interested. If you don't have 35 minutes to listen to the interview and another 5 minutes to read his blog entry (I've already done that for you), here's Marc's take on this move and its implications on the application server market:

Marc believes IBM is most likely NOT serious about this deal. Their lead WebSphere engineer did not know it was coming and said it was not going to affect his development plans for WebSphere. Marc's interpretation of this acquisition is that it's nothing more than a PR coup to create FUD and slow JBoss down while taking BEA out of the picture. All this with very little risk on IBM's part because, in his words: "Gluecode was a joke ... a dying/crappy company..." and it didn't cost IBM much anyway.

Fleury thinks that IBM's strategy doesn't make much sense, it is a "bait and switch" ploy where IBM wants to be OSS friendly until you are locked in and then they'll persuade you to upgrade to an expensive WebSphere/DB2 set up. He thinks this won't work simply because JBoss exists and JBoss has become good enough for the high end market rising up from its initial low end turf. His question is: why would people (customers and ISVs) pay for WebSphere if JBoss is FREE. My quick comment is to never underestimate what giants like IBM can pull off. I agree that BEA couldn't do it and that's why JBoss is squeezing them from the bottom up but IBM is a different beast. He also rightfully points out that Geronimo is not even on their competition radar today, they [Gluecode/Geronimo] still have to produce a 1.0 and a healthy community. Personally, I believe that, if (it's a big IF) IBM is serious about supporting Apache's Geronimo, it'll be on everyone's radar within 2 years. Look at what Eclipse (despite a late start) did to Sun's NetBeans.

Fleury also mentioned that, if anything, IBM's move confirms their fear of JBoss; it also validates that OSS is here to stay, it works and it's safe.

David Berlind asked Fleury a crispy question: why would IBM buy an insignificant OSS player [Gluecode] when they could have bought you and reached 67% market share? [The 67% figure comes from a BZ Research study that had JBoss at 34% and IBM at 33% followed by BEA at 27%]. Fleury's response was that he was not even approached because IBM is probably not serious about a genuine open source strategy; in addition to that "IBM doesn’t like us". This argument is definitely not convincing to me, people love you when you can double their market share. Other possible acquirers are obviously CA or BEA which could use a #1 spot (~ 60%) and become respected vendor and a force to be reckoned with, as it used to be.

Marc concluded that they were not interested in any acquisition at this point (I don't buy that, I think that it's only a matter of how much) and that he was going to have fun watching this movie unfold and enjoy being under the spotlight as a result (or side effect) of this "IBM PR coup".

Where do you think JBoss will be in 3 to 5 years? Dominant as an independent infrastructure player? Dead crushed by IBM's might? Or acquired by a much bigger vendor like HP, CA, Oracle or Novell? I am personally going for the latter. How about you?

Another interesting question: Is IBM going to force Sun into opening (in a FOSS meaning) Java, something they have been pressuring Sun to do forever?

With Geronimo's crew becoming IBMers and some of them are leading the Harmony project (the newly announced open source J2SE project), is IBM positioning itself to gain significant control on the Java platform despite Sun's resistance? You tell me.

Apple ditches key open source ally

A couple of years after Apple selected the KHTML rendering engine as the foundation of its Safari web browser, Apple decided to build its own version called WebCore. WebCore is an LGPL open source project.

The group behind KHTML described their relationship with Apple as a "bitter failure". They were expecting much more contribution from Apple. According to KHTML developers, Apple engineers took a less "pure" approach to fixing bugs, applying patches that were not good enough to include back into the KHTML code base.

"In fixing one problem, they were breaking a whole bunch of other things. Apple developers were focused on fixing bugs in such a way that we could not merge them back into KHTML. Those fixes were never an option for us." said Zack Rusin from the KDE, "In fixing one problem, they [Apple] were breaking a whole bunch of other things. Apple developers were focused on fixing bugs in such a way that we could not merge them back into KHTML. Those fixes were never an option for us."

"As long as they needed us, they used us, but when they gained enough knowledge they had no reason to keep sending us reviews and patches," Rusin said. "At a certain point they decided it was a waste of time for them, and at that point the communication just stopped...We had hopes that they would pour resources into KHTML. But that never happened."

Apple is not abandoning OSS completely; the company's Mac operating system is based on the Darwin open-source project (BSD-licensed).

For more on this story, read this article by CNET's Paul Festa.

My question (for those who know the answer) is: why is Apple developing a browser at all? Why not pour all their Safari resources on making the already popular Firefox fantastic on Mac OS?

IBM continues to strengthen its OSS-friendly image

This morning IBM announced the acquisition of Gluecode Software an open source application infrastructure company based in El Segundo, CA. Gluecode was founded in 2003 and received $5 million in March of 2004.

With this move, IBM solidifies its position as the leading supporter of open source software. IBM is a solid contributor to Linux with hundreds of developers on staff working on the open source operating system; it also backed the Apache Web server project and recently donated its very popular IDE framework Eclipse to the open source community.

It's interesting to note that Gluecode sells support services for Apache's application server Geronimo which is at the heart of the Gluecode stack. IBM's answer to the concern over the overlap between Geronimo and WebSphere is that Gluecode and Geronimo are not competition; they are door openers to enter new markets (companies or departments with less than 1000 employees) for which WebSphere is too big, too sophisticated and costly. "This widens our market," said Steven Mills, IBM's SVP for software. "Some customers may later move up to WebSphere if they choose to go in that direction." Chet Kapoor who joined Gluecode from BEA where he was the executive in charge of Quicksilver said: "We're very complementary to the strategy that WebSphere has now. They sell software at the high end, and we have a support and subscription business model for small and medium-sized businesses and departmental projects in big companies." Kapoor's career as a CEO will have lasted a couple of days as the news was announced on May 6, 2005.

A couple of random thoughts that I would like to get your opinion on:

- BEA continues to loose key people and struggles to find adjacent markets where they can grow, and my personal opinion is that they will go belly up unless somebody picks them up (Oracle maybe?)

- Did IBM (or the WebSphere group) feel too much heat in the SMB market from JBoss and decided to gobble up a start up to hurt JBoss by making Geronimo better and more popular?

- How important was Chet Kapoor in IBM's decision to acquire Gluecode? Chet knows so much about BEA and BEA's products (still today WebSphere's main competitor).

- This is great news for start-ups like Spikesource which is pretty much in the same business (now validated by IBM's acquisition) as Gluecode: providing enterprise service and test suites for open source stacks. Who will pick up this young Kleiner backed start-up? BTW, who's going to acquire Orbeon?

Please share your thoughts.

[Update:] Gartner published a research on this acquistion called Offense Is the Best Defense: IBM Leads Users to Open Source.

Open source Java - Hallelujah!

My day started off with some good news: The Apache Software Foundation announced today that they are kicking off a project called Harmony that brings together a group of experts (including people from Kaffe and Classpath) to come up with an open-source version of the next Java 2, Standard Edition (J2SE) – aka Java 5 aka Tiger. My personal belief is that there is an obvious need for such an initiative for Java's own good. The Java community and some of Sun's key partners (on the Java front) such as IBM have been pressuring Sun to open source Java.

In Harmony's FAQ, Geir Magnusson (in charge of Geronimo ) explains that they could only start this project now because Sun changed the licensing rules in the JCP to better accommodate open-source software.

Sun, traditionally a very strong ally of ASF, seems a bit confused after this announcement. It's VP and fellow Graham Hamilton expressed some doubts about the need for Harmony in his blog : "Personally, I am not entirely sure if the world really needs a second J2SE implementation…" He adds: "Personally, I am very curious about how the Harmony project will work out - creating a full scale implementation of J2SE is a mammoth task, as the Sun J2SE team knows only too well.” In other words he is wishing good luck to the Harmony group, their task will be overwhelming and he is probably right. Anne Thomas Manes, from the Burton Group seems to agree with him as well, she thinks it will take a long time to build a clean-room J2SE implementation, "because it's a very big piece of code."

Another challenge facing the Harmony folks is to make sure they are extremely careful about IP issues. Apache's Geir Magnusson concludes his FAQ with: "Historically, there has been wide exposure to VM and class-library-specific source code that is the property of Sun Microsystems as well as others... We wish to make every effort to ensure that the licenses and rights of external projects and efforts are properly respected. To that end, we will explore additional ways to work with the Apache Incubator to ensure that all IP is carefully monitored and tracked as it enters the project."

If you would like to know more about Harmony, Geir will be speaking at JavaOne 2005.

A basic architectural diagram of Harmony is available at: http://people.apache.org/~geirm/harmony.jpg