While Sun talks a lot about its open-source plans for Solaris, you are not giving us specifics. What are your plans in that regard?
Sun has contributed more to the open-source community in terms of lines of code than any other organization other than [University of California at] Berkeley. So, we are very familiar with the dynamic of the open-source community and what works and what doesnt.
But we are overburdening the words "open source." Theres a spectrum of licenses that apply to a variety of different contexts and applications. So you cant just ask if it is open source or not because the source code to Java is available; the bug database is available. But Java restricts you from mucking with the core concept of compatibility.
So what are your license plans for the open-sourcing of Solaris?
There are several licenses that may be appropriate in the open-sourcing of Solaris. While this is still a topic of debate inside of Sun, what is no longer under debate is whether we will open-source Solaris, which was a source of debate for a while. But what happened in Linux has really proven to us that having more code available to the community just brings more developers and customers.
Now that customers appear frustrated with Red Hat [Inc.]s licensing and their level of accountability around their products, there are all sorts of opportunities for us to deliver a better license that addresses a broader community and still opens more market opportunities rather than somehow threatening our economic model.
What about those open-source licenses currently available? Will they meet those goals?
The debate in Sun is now about the community model and the government model that you use to manage its evolution. The license is really a derivative of that conversation because if you pick the GPL [GNU General Public License], you end up with one set of issues and artifacts that arise, while if you pick the BSD, you come up with another issue.
So, what are your plans moving forward around this?
Let me divorce what the license model is from the objective, which is to broaden the appeal of Solaris to a growing number of new developers, as well as developers and customers really dissatisfied with Red Hats behavior. The other audience is those customers who want to have access to the source code. Truth is, we already deliver the Solaris source code to most customers who want it, but for some of them—especially governments, who are politically sensitive—we want to give them the assurance they have an alternative with us.
Thats why open-sourcing Solaris makes sense to us, as it broadens the community and brings more people in. Now, which license we use to go and pursue that is almost entirely orthogonal, as open source has a variety of different meanings. ... So which license we use is secondary to what we are trying to accomplish in the community. The third leg of this, which is in fact the more difficult in terms of how we go make progress on this, is, What is the community and engagement model we use inside of Sun to ensure that we can welcome contributions from others?
When can we expect a decision on this?
For us to get to the specifics of hosting the code and inviting the community to begin opening projects is going to take us probably six to nine months to flesh out in its totality. We have also begun really aggressive discussions with a diversity of developers in the Linux community and customers who want the ability to modify and enhance the code base themselves.
You are very critical of Red Hat and IBM. Why is that?
Red Hat has pursued a license model that has conferred a really big problem on IBM, which is that the license they have elected to use in no way, shape or form ensures compatibility across a community. ... One wonders why IBM is investing these theoretical 500 developers to ... work on Linux when it appears that what they are doing is adding value to Red Hat, who is then taking all those resources and putting them into building a richer and richer stack that is increasingly competing with IBM.