Blog Host: Bernard Golden - CEO, Navica READ ENTIRE BIO

I got a clue this week. I had coffee with a senior executive of one of the largest software companies in the world. He bemoaned the fact that, of a portfolio of over 80 products for which he's responsible, only about a dozen produce all the profits for his division. Many others, some of which are localizations of the base product into other languages, represent nothing more than a cost sinkhole. They'd love to jettison these unprofitable products, but, due to customer commitments, are required to carry them forward, release after release. His company doesn't know what to do, but knows that sometime soon they'll have to do something, because their company profitability is shrinking. Sooner or later they won't be able to carry the burden of maintaining these products.

By contrast, SugarCRM released its 1.0 product this summer. Within a few weeks, it had been translated into a half-dozen languages -- by users of the products. It costs SugarCRM nothing to make a product available in languages other than English.

In my experience, when two business models clash, the one that is more cost-efficient is the ultimate victor. I think I know how this particular struggle will turn out.

This is my last blog entry. I hope you've enjoyed reading the entries as much as I've enjoyed writing them. If you have additional questions or comments, you can forward them to the Expert Answer Center or to me directly.
Posted by Bernard Golden Why does open source foster innovation?
18 NOV 2004 11:45 EST (16:45, GMT)
One of the interesting questions about open source is "Why does it foster innovation?" Certainly the availability of open source components in an open source stack seems to make additional open source products happen very quickly. The growth of open source is incredibly dramatic. SourceForge, which is the largest repository of open source projects, has added over 20,000 projects this year. It's not just the fact that there are more projects -- open source is invading new beachheads, with new products being released in many different technology segments.

The contrast with Microsoft is striking. Its Longhorn release seems to be ever-receding and has recently been pared down to make a 2006 release date. Not to pick on the company, but its track record on Longhorn is terrible.

However, explaining why open source delivers so many products is difficult. After all, it seems like open source has many disadvantages. Many of the developers only work on their project part time. They're scattered all over the globe instead of collocated. They're often short of resources like equipment. And yet open source is churning out products. How?

It isn't source availability. While source availability is great, the vast majority of users and developers never look at the source code of their products. I would go so far as to say that most of the developers of a product that integrates with a second product never look at the source code of the second product.

It isn't marketing clout. Open source projects are notoriously poor, so they can't spend money on convincing other projects to integrate with them.

I think the very things that look like weakness are actually the reasons open source is being so successful. Because Microsoft has an integrated stack of products, they can't release one product without all of their other products being ready to go at the same time. This makes the last product to be finished the gating factor for release date.

By contrast, open source projects are developed in a disconnected, anarchic fashion. Because you never know what products may integrate with yours, you have to make general integration mechanisms available. Not needing to worry about other products allows you to finish yours as soon as it's ready. Its early availability lets other teams quickly take advantage of it to create their own products.

It's tempting to analogize the closed stack model and the open source model to the two economic systems represented in the Cold War. The command economy was eventually buried by the laissez faire capitalist economy. I'm not sure this is a good comparison, but it sure seems like open source is bringing products to market much more quickly than the commercial model. It's something to think about.
Posted by Bernard Golden What do users want?
17 NOV 2004 17:16 EST (22:16, GMT)
I attended the Software Development Forum's Open Source Conference today. The SDF is a Silicon Valley-based organization devoted to helping high-tech entrepreneurs start companies. Their conference had a number of panels discussing the future of open source, primarily from the perspective of startups focusing on open source. You can see the agenda and panels here.

There was lots of discussion and many assertions about what users of open source want. Some said end users want cheap software; other maintained that open source wouldn't really be adopted until there was a support company offering end-to-end support of the entire open source software stack.

One point made a number of times is that open source offers the opportunity for software companies to achieve distribution at a vastly lower cost -- instead of spending 50% of revenues on sales and marketing, the product can be put into users' hands for perhaps 5% of revenues.

I'm not sure what I think about the assertions about what users want. One thing I am sure of, however, is that the new open source model of a software company doesn't just get rid of expensive salespeople. The new model also won't be able to afford all the folderol we associate with enterprise software companies: the offsite seminars, the analyst tours, the glossy brochures, the expensive PR campaigns. I'd like to hear your thoughts on whether the new model of software company is something you're ready for. Let me know.
Posted by Bernard Golden Strategy vs. tactics
16 NOV 2004 06:46 EST (11:46, GMT)
One of the themes I speak about a lot is how open source is shifting from a tactical decision to a strategic decision. Whenever I speak to audiences of experienced open source users, I get lots of head nodding and enthusiasm. I get less enthusiasm when I talk about the implications of that shift.

Many open source advocates are pleased that open source software is now being considered for mainstream production systems. They're happy that open source will move out of the "good for internal use," less important systems. They're particularly happy that more money will be available for their efforts, as production systems usually are where the bulk of IT budget expenditure goes.

However, when I start speaking about the new processes that will need to be put into place for open source-based systems, I get a lot less head nodding. Talking about how training and documentation plans will need to be put into place seems pretty darn dull. There seems to be an expectation on the part of many open source advocates that the same free-wheeling methods of development and implementation will transfer across to the new production systems.

In a phrase, "it ain't gonna happen." If you are a senior manager responsible for running production systems, you are going to want to see plans and processes in place to deal with all phases of the project -- development, implementation and operations. Metrics, Gantt charts and project roadmaps are all part of the big-ticket project reality. Thinking that it will be different because open source is part of the equation is just setting yourself up for disappointment. Far better to recognize that upper management is going to want to see these traditional type of project plans and metrics and to begin figuring out how to apply them to open source-based projects.
Posted by Bernard Golden Cheap and easy: A winning combination
15 NOV 2004 05:32 EST (10:32, GMT)
One of the main criticisms of open source is that, while it tends to be simple, it "doesn't scale." The implication of this criticism is that if a piece of software can't handle the largest potential loads that might occur in any production environment, it's not a viable choice. This criticism is often voiced in a condescending tone, implying that open source doesn't deserve to be mentioned in the same breath as "real" commercial products.

Frankly, I don't understand this attitude. If we only purchased products that could handle the most extreme conditions, we would only drive Ferraris, 16-wheel semis or Hummer SUVs. Instead, most of us opt for more general-purpose, less expensive alternatives like Ford Explorers, Honda Accords and the like. And, by the way, there are way more Accords on the road than Ferraris.

Moreover, using a less expensive alternative offers more flexibility and can speed adoption. To use an example more germane to technology, let's look at networking. Today, the entire world runs on TCP/IP, but 20 years ago TCP/IP was considered an inferior contestant to the International Standards Organization's Open Systems Interconnect (OSI) standard. Endless articles described the OSI seven-layer networking model with examples of how it would revolutionize connectivity. Everyone expected it would dominate networking; meanwhile, TCP/IP would be discarded as an doomed evolutionary path that lost out in the natural selection of networking protocols.

The reality, of course, turned out much differently. While TCP/IP is not as elegant, it was freely available, easily understandable and, perhaps most importantly, put into actual implementations while the OSI model iterated through endless discussions and failed interoperability attempts and promised that it was the "wave of the future."

We're fortunate the story turned out as it did. We are the beneficiaries of the triumph of TCP/IP. With a cheap and easily available standard, companies put out inexpensive, innovative products -- products like DSL and 802.11x wireless access networking. This is not including actual open source products that take advantage of the ubiquitous availability of open source -- products like instant messaging and VOIP. I hate to think of how difficult it would have been to achieve these kind of things in an OSI world.

Similarly, the easy availability of cheap software encourages innovative experimentation and distribution of useful applications. Interesting products can be turned out that never would have seen the light of day if they required expensive infrastructure underpinnings. Cheap and easy may not solve every problem, but makes it possible to solve whole new ranges of problems. Keep the Accord in mind when thinking about open source.
Posted by Bernard Golden Where will open source go next?
12 NOV 2004 05:48 EST (10:48, GMT)
If you read the trade press these days, you see lots of articles about Linux. You see a few articles about MySQL. An occasional mention of JBoss. And, every once in a while, an offhand comment about some other open source product like Snort or Nessus. It would be easy to think that the open source world is made up of five or six products -- maybe 10 or 20, but certainly less than 100.

The reality is vastly different than that. If you go to SourceForge, there are 90,000+ open source projects. Yes, some of them aren't worth the bits they take up, but I am astonished by the breadth and variety of the open source programs available for use.

To give a couple of examples:

1. I was doing a search on SourceForge for estimation software relating to software projects. In the search results were several open source construction estimation projects. This is an area I know nothing about, but I was amazed that someone had taken it upon himself to create this product and distribute it under an open source license. A criticism often voiced about open source is that it is very IT infrastructure-oriented. You can't get much further removed from IT infrastructure than estimating a kitchen remodel!

2. I was contacted by someone who had read my book and was trying to figure out how to apply the Open Source Maturity Model in his business, which is installing open source PBX systems. We had a good discussion about the future of open source PBXs vs. the traditional proprietary software products. He is busy putting open source-based products into non-tech office environments with a value proposition that his system is less expensive than the alternatives. (He very kindly offered to put a review of the book up on Amazon, which you can take a look at for a nuts-and-bolts business perspective on open source).

This blog entry brings my first week in the Expert Answer Center to a close. You've probably gotten a sense that I believe that open source carries enormous economic implications for the creation and use of software. The two examples described above give just a taste of open source's impact. I hope you've enjoyed the blog so far. I look forward to next week and your questions.
Posted by Bernard Golden Firefox: The 30-second migration
11 NOV 2004 05:38 EST (10:38, GMT)
One of the biggest complaints you hear about open source is that it's not very user-friendly. Open source works well, but isn't as easy to install or configure as it could be. There are lots of reasons why this situation exists, but the received wisdom is "Open source: Good but not easy." This reputation causes lots of organizations to defer migrating to open source because they think the transition will be difficult. However, our recent move to the Firefox browser gives me hope that open source projects are starting to pay attention to making migration easier. Let me share our experience.

We decided to migrate our users to Firefox from IE, due to security concerns. We had held off doing so due to a combination of laziness and concern that the migration would not go very smoothly; in particular, we were concerned that our bookmarks would not be transferred to Firefox. Now, before anyone writes and says, "Firefox automatically transfers bookmarks," let me say that we were aware that it claims to do automatic transfers; however, we had a different bookmark setup and weren't sure that Firefox could handle it. Let me explain further.

We've had problems in the past when doing the inevitable reinstalls of Windows. Because Microsoft insists on treating IE as part of the OS, it keeps a lot of IE's data, including bookmarks, in areas that get wiped out during the installation process. Because we found losing our bookmarks so frustrating, we used the very handy TweakUI utility to place our bookmarks in a different location than the default place the Windows install program puts them. This enabled us to avoid losing our bookmarks when the time came to reinstall Windows due to registry corruption, application failure or mysterious refusal to boot up.

We assumed that Firefox was programmed to look in the default location for IE bookmarks and transfer them across, but that it would fail to search for bookmarks placed in other locations. Nevertheless, we decided to do a trial installation just to see what would happen. It was amazing. Firefox found the bookmark file deep in the bowels of user storage and made the bookmarks available immediately. We were up and running with Firefox literally within 30 seconds from the time we hit the install button. It was fantastic!

This experience gives me hope that open source project teams are incorporating usability and migration functionality into their project plans. In the past, these items have not been included, mostly because the user population was extremely technical and capable of addressing these issues directly. With the spread of open source, new types of users are coming to the products with less technical skill. I've been hopeful that the projects would take this new user type into account; our experience with Firefox indicates that this is indeed happening.

This is a really positive trend for open source, which will speed its adoption. I commend the Firefox team for its work in this regard.
Posted by Bernard Golden My interesting open source day, part 2
10 NOV 2004 05:46 EST (10:46, GMT)
In yesterday's blog, I wrote about my discussion with Sam Greenblatt, Computer Associates' point man for open source. CA's decision to open source the Ingres database limns one potential path forward for open source -- corporate technology providers divining new business models that incorporate open source as part of their portfolio of products and services.

On the same day I spoke with Sam, I had dinner with Richard Stallman and a number of friends. While I had never met Richard, I had certainly known of him for a number of years. Before there was an open source movement, there was Richard.

For those of you unacquainted with Richard's efforts, he began working with Unix a number of years ago. Early Unix users freely swapped home-grown programs to enable one another to better administer their machines and build applications. Perhaps inspired by this easy sharing, or turned off by the financial grasping that started shortly after Unix commercialization, Richard concluded that software -- all software -- should be available for free. Furthermore, there should be no restrictions on how software could be used. Finally, the software should be licensed in such a fashion to preclude it being transformed into commercial products.

From that inspiration, he started the free software movement. Based on the principles I outlined in the previous paragraph, it began delivering software distributed with the Gnu Public License, often referred to as the GPL. However, a young Finnish student also began delivering software that became Linux. It caught on rapidly. A few years after this, a group of people met to come up with a more palatable term for this type of software, concluding that "free software" was confusing. Their new term was open source. The rest, as they say, is history. Open source software creation and use has skyrocketed over the past few years, far surpassing the availability of free software.

Richard Stallman has not disappeared into the night, however, He continues to pursue his dream with single-minded purpose. He frequently appears in technology stories and has found himself in the center of a number of controversies. For example, he protested loudly against the introduction of tracking badges in a new building at MIT, where he works. He is no shrinking violet.

I definitely looked forward to meeting him.

I wasn't sure what to expect. I've had plenty of experience over the years interacting with ideologues, having gone to college at UC Santa Cruz. Many of these individuals seem to feel that a conversation that doesn't end with a heated standoff somehow is lacking substance. I thought a food fight over the dinner table was a definite possibility.

Actually, it didn't turn out like that at all. Due to a mix-up, I ended picking Richard up at the house he was staying at. He gave me a copy of his latest book, inscribed "Happy hacking." We had a very nice conversation on our way to the restaurant. Dinner passed quite pleasantly, with nary a raised voice.

However, I learned that his conviviality by no means indicates a lack of passion for his ideals. He is a very firm believer in free software. You could see him shudder when the term "open source" came into the conversation.

His perspective is simple: All intellectual property should be free and freely distributed. The laws of the country should enforce that policy. Richard doesn't seem to be phased at all that what he sees as the illegitimate offspring of free software -- open source -- is getting enormous traction or that free software is less and less seen as an important movement. He will continue to advocate his ideas no matter what the trend is. I came to think of Richard as a genial zealot -- pleasant, but passionately committed to his beliefs.

As you can imagine, a day that contained both Sam Greenblatt and Richard Stallman is a very interesting day, indeed. I have no doubt that new business models will sweep through the technology industry, reflecting the unstoppable momentum of open source. As vendors begin to accommodate open source, the open source movement will continue to morph. Sam Greenblatt represents one possible future for open source, while Richard Stallman is far more likely to be seen as a beginning rather than a destination for the open source movement.
Posted by Bernard Golden My interesting open source day
09 NOV 2004 06:07 EST (11:07, GMT)
Writing a book is a broadening and inspiring endeavor. As I told people about the project, they'd say, "You should talk to so-and-so. He's a real expert." I met and got to know many, many people that, absent my book, I never would have had reason to speak with. And that's what led to my interesting open source day.

In a single day, I had the opportunity to speak with Sam Greenblatt, Computer Associates point man on open source, and Richard Stallman, considered by most people to be the father of the free software movement. It was, as they say, an interesting day. In today's entry, I will discuss my conversation with Sam; tomorrow I'll share my meeting with Richard Stallman.

Sam Greenblatt
I spoke to Sam to discuss CA's decision to offer the Ingres database under an open source license. Ingres was a high-flier in the early database wars, but fell on hard times. CA purchased the company, but Ingres fell behind the market leaders and eventually became an "also-ran." It was kind of a shame, because Ingres was always considered a top-notch product from a technical perspective. On the other hand, from a market perspective, CA probably saw a future of continued investment for little return.

What Sam told me was that, alongside this product-specific assessment, CA itself is changing course. They are downplaying software licenses and focusing on a subscription/support revenue model. In some sense, their corporate direction fits quite well with the realities of the Ingres market. Consequently, they decided to convert Ingres into an open source product, make it available for free download and offering services to users who wanted a more formal support mechanism.

I enjoyed speaking with Sam, because I believed (and believe) that open source will transform the software industry. The old model of big, upfront license fees is all but dead. Technology companies, if they wish to survive, will have to mutate their business models to more of a pay-as-you-go service. It's a cliché to say "software is becoming a service," but that statement captures part of the new software landscape. Another part of the landscape is that service businesses are judged by their customers by very different criteria, and the transition will be extremely painful for software vendors; not all of them will make it. Furthermore, CA's strong endorsement of open source reflects the changing nature of the open source movement. As the decision to use open source becomes more common -- and more strategic -- the big technology vendors have begun to support it. Their involvement in open source has, inevitably, changed its nature.

Big technology vendors and big technology customers think different (if I can steal a line from Apple). They focus on issues like "How do we train a worldwide sales organization to push this offering?" and "How will we roll this out to our vertical industry segments?" As the big vendors and customers barge into the open source cocktail party, the conversations you overhear will be quite different. Not every open source participant will be happy about the gatecrashers. However, it's inevitable. As new technology movements spread, the original band of enthusiasts (typically passionate, but not very corporate) are joined and even shoved aside by a mob of smoother, but less enthusiastic, corporate employees. I've seen this happen in two or three technology movements and see it beginning with open source. The animal we call open source will be quite different in two years -- bigger but more tame.

Tomorrow: my dinner with Richard
Posted by Bernard Golden How I got my start in open source
08 NOV 2004 06:05 EST (11:05, GMT)
Since I will be blogging over the next two weeks on the topic of open source, I thought a good first posting might be to share with you why my company first began working with open source, and what convinced us to focus on open source as our business strategy.

We were working on a project for a large health care foundation, creating an integration between their external Web site and an internal CRM system that they used to distribute online newsletters to their subscribers. The need for the integration had been recognized late in the project, which meant the budget for it was pretty limited. In our search for an inexpensive integration mechanism, someone recommended that we consider using a Perl module SOAP::Lite, which provides Web services functionality.

I was pretty skeptical about open source, for all the usual reasons you hear when discussing open source -- The developers must be flaky. How can you trust software that has no owner? Where do you turn for support? -- in short, all the usual suspects! However, I felt we had to give serious thought to the Perl option due to the budget constraints we were operating under.

What happened during the course of that project hit me like a ton of bricks. As the different groups worked on their parts of the project, we all ran into problems -- the product didn't seem to work the way we thought it should, the documentation wasn't clear about how to configure the product and so on. All of the usual stuff of a software project.

However, while all the groups faced the typical challenges of a software project, our experience in moving forward was vastly different. As our group ran into problems with the Perl stuff, we posted questions to the appropriate mailing lists. We would often get responses within 20 minutes, posted by people located throughout the world. This enabled us to move forward with our part of the project very quickly and efficiently.

Meanwhile, the group working on implementing the CRM product also ran into problems. Again, these were the typical problems associated with any project implementation. The client had spent a ton of money on the product and a lot of money for a support contract. However, the support they received was terrible. It took the vendor two days just to accept the first support call, which was, "Do you have a support contract? I can't find you in the database..." Once they got that straightened out, they got someone answering their call who knew less than they did about the product. Eventually, they got through to someone who understood the issue and assured them it was fixed, in a later version of the product, which they would need to purchase if they wanted to solve their problem. As you can imagine, their part of the project lagged significantly -- to the client's detriment.

At that moment, I knew this thing called open source was going to shake up the world of information technology. I resolved to understand it better and see if it could benefit all our clients. As I researched open source further, I concluded that it would soon come onto the radar for all IT organizations and a guide to successfully begin using open source would be very useful. Thus, my book Succeeding with Open Source.

I also concluded that the fundamental economic trends within IT organizations would force them to begin using open source. (You can learn more about my perspective on this in the webcast I did last week called "Why your future depends on open source." It can be viewed here.)

Well, this has been a bit long-winded for a blog posting, but I hope you've found it useful as an introduction to the next two weeks. I believe that open source is becoming a strategic decision for every IT organization. Learning how to use it effectively offers enormous benefits. Dismissing it is wrong-headed, and, more important, will expose you to serious competitive disadvantage in the future.

I look forward to the next two weeks and hope you do, as well. Please forward questions that I can answer for you.

Tomorrow's blog: my interesting open source day.
Posted by Bernard Golden