Open source can of course mean many things and there is no clear agreement on a definition.  The Wikipedia entry as a POV notice on it for example.  There is also an important difference between the concept of open source in general, and that of open source software in particular.  In respect of software we have the following definition from OSI.

Open source is a development method for software that harnesses the power of distributed peer review and transparency of process. The promise of open source is better quality, higher reliability, more flexibility, lower cost, and an end to predatory vendor lock-in

However in wider use the term also means the creation of shared content, which is freely available to all generally under the various legal structures available under the creative commons.  Now there are ideological approaches  which argue against any concept of ownership or control.  There are also general phrases such as open source intelligence which simply indicate that material is publicly available.

The arguments for open source (generally) and open source software (specifically) are in the main a mixture of the pragmatic and the ideological.  The one thing that is clear is that all successful open source movements make money somewhere in the system, and until the fall of capitalism and/or the availability of unrestrained resource (neither of which are likely) this will remain the case.  On the side of pragmatism the concept of large numbers of people testing and working on something is advanced as improving quality, and the same is an argument for increased likelihood of innovation.  At the same time, such an approach also means that many views are taken into account, which can also lead to bland conformity with the orthodoxies of the present, or of the interest group who engage; become an active editor in the Wikipedia if you want examples of this.

Open source software models in the main switch from license revenue to service revenue.  You don't have a lock in on the software (anyone can access it) but attempts to lock in on service fees (installation) or by licensing specific applications or bundles with proprietary software.  Some models seek to dominate a space with huge Venture Capital investment, and then use the platforms they create to make money through advertising and other services.  This is not always visible, frequently not transparent and is anyway subject to the needs of the VC community.  Even Angel investmentors are not immune to commercial gain.

Making choices, but money has to come into the equation somewhere

Now there is a hidden assumption here, namely that one needs to make money.  I will freely admit to getting irritated with academics and public servants who from the security of a monthly salary, pension and other benefits criticise those of us who, in order to create something new, have had to work for ourselves or create start up companies.

Anyone doing this faces a series of choices.  In Cognitive Edge we decided that we would retain license fees on the software, but we would not restrict access to that software other than a requirement to attend a training course.  Services are thus "free" and unrestricted but the software is not.  Other groups would reverse that position.  Neither is right or wrong per se.  However to say that there is a contradiction between open source methods and proprietary software is to display a woeful ignorance of the open source movement and economics in general.

There is also a question of focus.  Sometimes, to establish a novel or different way of doing things it is necessary to keep a tight focus.  Once a core way of operating has been established then it is possible to change the nature of models.  For example making API's open so that other people can develop software which links and connects is a logical and allows a broader community of developers to engage.  However to get there requires investment and time.  It doesn't happen overnight.  Mixing open source and proprietary models of software development is also increasingly attractive, as a product matures and its user base increases.

If you are developing software against a known need with a large body of people who already understand the principles then a "pure" open source method might work.  However something novel, where the concept as well as the software has to be sold requires a very different approach.

Other business models give everything away on the basis that money can be made through advertising.  This is a common one, but increasingly subject (as Doc Searls points out) to control by a reducing number of players.  I previously referenced the danger here if you look at history.  I am not sure how long the advertising model, with its high wastage can survive, it shows all the sigh of a bubble, and when that bubble bursts we are likely to see a field dominated by a limited number of strong players with large capital bases.  Often the business models here are not transparent and as I said earlier I think its the transparency of the business model that counts, less than its nature.

There is also the question of time and place.  I love the concept of Linux, partly because I don;t like Microsoft.  However I use OSX because I only want to use an operating system as a tool to achieve other purposes, I am not interested in operating systems per se.  We are undergoing an period of extreme turbulence in technology coupled with economic, political and environmental uncertainty.  There is no guarantee that current models will persist or scale.  Sometimes it is necessary to create a proprietary approach to make progress in a unique field in whole or in part, and wait for critical mass to buy into the approach.

So its not a simple open source good, proprietary bad position.  It is far more complex, with many a difficult choices to make.  In Cognitive Edge we chose not to restrict access to the methods (anyone can be trained and then contribute), or (other than through training) to the right to use the software.  Other than for charity cases or future research we do not discount or differentiate fees - we keep a level playing field for all practitioners.  Critically we also, through revenue share, allow those practitioners to create sources of income that will free then from the constant tyranny of utilisation.  By providing tools we give them a means to survive outside the structures of large consultancy firms.  We also provide people in government and industry with access to that network and its tools to again reduce the dependency on manufacturing models of consultancy.  Not only that we have managed all of that todate without selling a large part of the company to a VC who might well have very different objectives.

So I don't take offense at remarks such as the above mentioned tweet, it's a simplistic comment based on a naive perspective from a protected position on what it takes to create software tools these days, particularly ones that challenge management orthodoxies.  The one thing I can tell said Tweeter is that IBM never got it, and never will.