Sun is getting into the open-source business with Solaris, but will this automatically translate into better sales?
"But what about those people who can't tell a heap from a stack, and certainly couldn't design a safe linked list de-linker? Who do they trust to make that decision for them?"
Open-source vs. closed-source
Pure Unix vs. Unix-like
Better performance vs. better security
All of these issues have been discussed on blogs and addressed by columnists over the last few months. With Sun Microsystems ratcheting up their considerable PR strength for their upcoming open-source Solaris, and with a bit of a rock throwing contest developing between Sun and the rest of the open-source OS world about which is the better open-source vision, my crystal ball is still a bit hazy with what the future will look like. I will go out on a limb and suggest that the developments of the next few months are going to significantly affect the "look" of the *nix world moving forward.
Sun is getting into the open-source business. Is this a good idea for them? They obviously believe that there is an upside to them giving out more information. Where does this belief that open-source is inherently better actually come from? Will they open-source everything? While the answer is unclear, the cynic in me remembers the lack of print support in the open-sourced version of StarOffice a few years ago. What will they do with Solaris?
An issue of trust
I believe the heart of the matter comes down to trust, so let's start with an analogy.
Most avid news readers will have noted that the big pharmaceutical companies are having a little bit of a trust issue with some recent drugs to treat depression and arthritis pain. The problems are related to drug safety, clinical trials, and the approval process of these drugs. In response to the allegations that there wasn't adequate disclosure during testing, or that the US Food and Drug Administration approved these drugs without appropriate oversight, there are increasing calls for a public clearinghouse that will document the results of clinical trials for all drugs.
Now, I don't know about anyone else, but I don't feel qualified to judge the quality of clinical drug trials, or interpret the results, so why should this clearinghouse of information make me feel better? Why is the mere existence of this website a good thing? Am I able to judge if I am at greater risk of a heart attack by taking a pain killer? Remember, I am essentially a naive end-user of these drugs.
The reality is that, most of the time, the end user doesn't have the knowledge to make an informed judgment.
Basically, I trust that someone who is competent (or at least plays a competent role on TV) will be able to make this determination for me, based on this public information. This mythical competent-someone is essentially a check-and-balance, or in an extreme case, a whistle-blower or end-user advocate. The million dollar question is, "does a public clearinghouse actually lower the risk of a bad drug making it to the public?" I certainly don't believe that it's a step forward for me to require a PhD in bio-chemistry to assess if I should take Advil or Tylenol for my headache. That's why there are experts out there - hopefully, independent ones.
In the case of the pharmaceutical industry, these calls for an "open-source" of drug trials came about because of a lack of trust in the pharmaceutical companies and the FDA in ensuring safety and efficacy. Compared to pharmaceuticals, the software world has precious few regulations to ensure effectiveness, safety, usability or anything else that might benefit a consumer. Consumers are forced to trust the authors of the software to ensure that the software does what it says.
The notion of open-source software came about historically for many different reasons (for those interested, the history of BSD and GNU would be good starting points), but in many ways, open source is now being marketed as a (essential?) contributor to security. The argument is that you can verify the code yourself, or fix it yourself if something really bad happens. But what about those people who can't tell a heap from a stack, and certainly couldn't design a safe linked list de-linker? Who do they trust to make that decision for them?
Do customers care?
Perhaps it isn't fair to ask "does anyone care?" After all, I care, and the volunteers in the Free Software Foundation care, the developers of Linux and the BSDs care. But do the consumers care? Do enterprise customers care? Is there a hue and cry from the great unwashed for open-source? Does Joe Q. Public distrust the closed source software from Adobe, Microsoft, or even a spyware company enough to buy or download an open-source alternative?
Let's put it another way; are people installing Mozilla or Firefox because it's open-source and they can trust it, or because they like tabs and pop-up blocking? If the answer is because they like the functionality of the tabs and pop-up blocking, then open-source as a marketing feature isn't much of a marketing feature. Rather, open-source is simply a way of doing business, one that doesn't mean much more to the buyer than what development environment the software was written in, or what source control software was used.
For a business customer, does Susan P. Executive care that her servers are running an open source system, one that a competent whistle blower might have verified is safe? Perhaps for Ms. Executive it is more important that she has someone to turn to if something does go really bad. If she wants large company to stand behind the software, why would she pick Novell or Redhat instead of Sun, SCO, HP or Microsoft? Or perhaps a better question is, why will she be more likely to pick Sun if they open source Solaris than if they don't?
Again we must ask, why is Sun contemplating open-source for Solaris? I don't know the answer to this, and like others, I've read the reasons given by Sun. But I just don't see a critical mass of consumers demanding access to the source code. Certainly there are niche markets where source access is important, but are they so locked out of those markets by Linux or BSD that they need to open-source Solaris? I doubt it.
What is open-source good for?
Now that I have spent most of the article undermining a frequently argued case for open source, why do I think open-source has value? I believe that we can learn a tremendous amount from what has been created before. Additionally, open-source or free-software has certainly built a lot of the infrastructure of our connected world. BSD, Apache, Linux, the GNU project, Perl, PHP and Python. The list of building blocks of computing today, especially when dealing with the delivery of information over the network, is impressive. And the change that this information delivery has brought to our world society is profound.
The value in open-source is in propagating knowledge, not in making computing safer or easier for a naive end-user or business consumer. To argue otherwise is misunderstanding the needs and wants of the customer. Sun may be able to turn their fortunes around, and indeed some of their technology for safer and high performance computing certainly looks very intriguing, but I don't believe that open-sourcing Solaris will make much of an impact for their customers one way or the other.