Interim
There are two approaches to promoting open source software: 1) avoid any use of existing non-free software, or 2) use some non-free software while free alternatives are being developed. The danger in the first approach is that people will not adopt the solution. The danger in the second approach is that free alternatives will not be developed fast enough because the proprietary alternatives are too compelling. Freespire takes the second approach for the most part. The first approach (an operating system that uses 100% open source offerings), is already well represented in the Linux marketplace, and Freespire salutes those who are representing that approach. Hopefully between their good work and the approach offered by Freespire, Linux will find its way to a much wider audience of mainstream users. (Freespire also supports the first approach with their 100% open source version.)
I think the Freespire FAQ author was very straightforward about this, but I thought of six other problems with the second approach.
- It provides benefits to proprietary software and technologies through network effects (and helps extend the first-mover advantage of some proprietary technologies by giving them an even greater market share).
- It may displace some development resources from free to proprietary software.
- It may make the distinction between free and non-free software more obscure to some users, and cause some users never to become aware of the distinction at all.
- It may persuade manufacturers that they don't need to document their hardware because they perceive the "Linux community" accepting proprietary software. Analogously, it may blunt the ability of Linux customers to send market signals to manufacturers that free software support is necessary.
- It blurs the distinction between technologies where a free software implementation is slow in coming (e.g. Flash, Java) and technologies where a free software implementation already exists but can't be widely commercially distributed some some jurisdictions (e.g. libvdcss) or where free software development will be deterred in the long run for legal reasons. The Freespire FAQ alludes to the risk of free software "not be[ing] developed fast enough" for some application, but doesn't discuss the prospect that legal considerations could prevent it from being developed at all.
- It makes it more difficult for people to understand why laws that burden free software development are a problem; "there is no Linux DVD player" was a much clearer message for many people than "Linspire's DVD player is not open source".
Richard Stallman has famously suggested that it's only legitimate to use proprietary software for the specific purpose of creating a replacement for that software. For example:
To develop GNU, I used Unix. But first, I thought about whether it would be ethical to do that.
I concluded it was legitimate to use Unix to develop GNU, because GNU's purpose was to help everyone else stop using Unix sooner. We weren't merely using Unix to do some worthwhile job, we were using it to end the specific evil that we were participating in.
This seems weird even to many free software advocates because it is so radically prescriptive not only of how software ought to be distributed but even of end users' apparently isolated choices. (Richard doesn't seem to be saying merely that it's wrong to publish, enforce licenses of, promote, or pay for proprietary software, but even that it's wrong to use it!) But you don't have to reach that question in order to see a problem with Freespire's approach -- you only need to consider the existence of network externalities. Among other things, network externalities mean that the installed base of people who can receive something affects the decision facing someone who wants to communicate with them.
