While we tend to have pretty good visibility into Sun, they do manage to sneak something by us every now and again. Such was the case with today’s surprise announcement that Ian Murdock is leaving the Linux Foundation to join Sun. I guess we shouldn’t feel too bad, however, as Tim says that the hiring was conducted in Ultra secrecy (although it would appear that the move has actually been in the works for a few weeks now).
In any event, given the multi-level significance of the news, I feel compelled to comment. Via a Q&A, of course.
Q: Before we begin, do you have any disclaimers to make?
A: Ian’s new employer, of course, happens to be a RedMonk customer, as are a couple of the members of the consortium he formerly called home. On a personal level, I’ve been fortunate enough to meet Ian in person several times and have a great deal of respect for him. Not solely for his community and technical contributions, which are immense, but also his approach and his rationalism. I can think of no higher compliment to bestow on him than the title of pragmatist, and from his comments it would seem that Bryan would concur with that assessment. In other words, I’m a fan.
Q: How about summarizing the news for those that haven’t yet seen it?
A: Sure. The short version is that Ian is leaving a CTO role at the Linux Foundation to take a position as Chief Operating Platforms Officer at Sun. While that title would seem to leave little to the imagination, Ian was actually somewhat circumspect about his actual responsibilities, and Tim indicated that the plans aren’t fixed. My understanding at this point is that the role reports up through marketing, as opposed to engineering, but that’s not atypical for Sun.
Q: Beyond the immediate implications in terms of roles and responsibilities, what are the wider implications to this hire?
A: First and most obviously, it’s a high visibility hire. Virtually everyone in the Linux community knows Ian, and will take note of this move. Much as Bray was noted for his work on XML and Phipps for his advocacy of Java within IBM, Ian is no ordinary technologist: he brings with him a high profile, and the attendant benefits will likely accrue in some fashion to his new employer.
Second, it’s a significant hire in a part of the organization that is in some flux at the moment. As the Register covered, some of the higher level Solaris executives – folks I know and respect – have recently departed the company. Add in the fact that due to a variety of factors, many of the folks that built the differentiating features that got Solaris back in the game are now working on things other than Solaris, and there’s cause for concern. Ian by himself is not the solution to the aforementioned issues, but the hire is certainly timely.
Q: I take it from the above that you think this is a good move for Sun?
A: I do. My actual response when someone from Sun pinged over the news this morning was, “how’d you manage that?” It’s not a move without risks, and I’m sure the responses will vary from community to community, but personally I think this is an excellent hire for Sun. I haven’t been able to determine yet whose idea it was, but kudos to that person. As Tim says, when you have the chance to hire someone like Ian, you do it and sweat the details later.
Q: What do you see as the primary risks to the hire?
A: They’re two sides to the same coin: Ian meets significant internal resistance and is unable to effect necessary change, or Ian meets resistance and chooses to boomerang out of Sun. Sun has long been a relatively political organization, even for its size, and the divide within the Solaris community in particular is – at times – quite wide. In short, the question to me will be can Ian win over or persuade the old school factions within Sun? The new schoolers seem to be presold. Rightly or wrongly, Ian is going to be viewed by many within the Solaris and OpenSolaris communities as the “Linux guy,” and that lens alone is likely to provoke some non-rational, emotional responses given the antipathy a fair percentage of members of the Solaris community have for Linux. Dealing with that will be a continuing challenge. Ian’s no stranger to such friction, of course, as Linux communities are not known for their tendency to hold hands and sing kumbaya, but there are significant organizational and governance differences between Linux and Solaris/OpenSolaris.
Q: Can you be more specific regarding the divide between the two schools?
A: The most visible example of the divide probably came in the form of the original decision to open source the operating system in the first place. Many engineers and executives had been agitating on behalf of an open Solaris for years within Sun, with little success. One of the difficulties, apart from the cost and effort required to open source such a sizable project, was the entrenched strength of those who opposed such a move. Eventually, of course, the new schoolers prevailed and we now have OpenSolaris. Despite the apparent and evident success of that project, there remain pockets within Sun and without that remain decidely old school.
Q: Is the open source vs closed source divide the primary point of contention between the new and old schools?
A: No. The tension between new and old manifests itself in a variety of other areas, including binary compatibility, choice of shell, installation experience, licensing, package management, userland tools, and so on. And frankly, it’s not unique either to Solaris or to Sun – it’s natural to have factions pulling in different directions. It’s just that within Sun, leadership changes at the highest levels have changed the dynamics between the factions quite dramatically in recent years.
Q: Do you view this as an offensive strike at the Linux Foundation?
A: No. First of all, Sun is, by virtue of its membership with the LSB I believe, a member of the Linux Foundation. Second, Ian will continue to chair the LSB. But last, this is about hiring the right person at the right time – not trying to damage another organization, despite some of the comments that have come out of the LF that dismiss Solaris as real competition.
Q: Do you think this is a mandate for change for Solaris?
A: Mandate? No. From the conversations I’ve had, the Sun folks are already quite clear on where they’re scoring a Needs Improvement on the report card. Recent chats with Solaris folks have indicated an impressive willingness to change, to adapt, to evolve. Simon, Tim, and many others across Sun are quite aware that package management, as an example, is a significant advantage for Linux at the current time. So no, I don’t think Ian’s hiring can be read as a mandate for change.
That said, however, I think it would be a mistake to hire away the founder of Debian and not drive significant change. A lot will depend on how empowered Ian is in his new role; from what I’m led to believe, he’s got support all the way up, but even with that organizational inertia and resistance can be difficult to overcome quickly.
Q: What do you anticipate Ian will be working on?
A: The expectation in most quarters – see Jason’s comment here or Bryan’s final comment here – is that Ian will be tasked with remedying the lack of package management functionality. That expectation, of course, is quite logical given Ian’s background, and I share it. I’ve argued on behalf of apt-get to Solaris engineers and executives on multiple occasions (more on a popularity than technical basis), and one would be foolish to read nothing into Ian’s hiring in that regard. Beyond that, I’m interested to see whether or not the GNU userland can make inroads into Solaris, and what Ian can bring to the table from a package standardization perspective.
Q: What would you like to see happen from this hiring?
A: In a perfect world, as Mark says, this hiring would result in some convergence between the two operating systems. One of my stated regrets regarding the original Linux Foundation announcements, in fact, was the unhelpful wedge some of the surrounding rhetoric seemed to drive between the Linux and Solaris worlds. Perhaps, with Ian’s move, that divide need not be permanent.