At the risk of sounding a little silly, I’m going to attempt to extend the analogy between Sakai 3 and Mac OS X that I made in my last post. The reason I think this exercise is worth trying is that the Mac OS X transition is a relatively clear and uncluttered example of a successful rearchitecture of a product with an installed base, and it provides us with a good model of what could happen with Sakai 3 if the initiative is executed well over a number of releases. Please understand that I am not predicting anything here. I’m just trying to model possibilities so that we can set a ceiling on expectations.
Let me start by acknowledging a couple of dysanalogies up front. First of all, as Michael Korcuska points out, deciding to change your personal computer’s operating system is different than deciding to change your institutional, user-facing enterprise application in that it is much harder to reverse the latter decision. I think that, as Josh Barron comments at the bottom of Michael’s post, the ability for institutions to pilot and to run up limited-use installations narrows that gap somewhat, but the larger point is still worth acknowledging. Second, and perhaps more critically, Sakai is a community-developed open source project, where Mac OS X is a corporate-supported private source product. This makes a big difference in the predictability of resourcing and therefore pace of evolution. Apple’s resourcing was highly predictable, where Sakai’s is not. On the one hand, positive responses to early Sakai 3 releases may attract new contributors to the community and accelerate development. On the other hand, because the current installed base is less likely to be enthusiastic about an early migration than new adopters will, and because said installed base is the main source of major code contributors who will drive the pace of Sakai 3 development in the early releases, there is a danger that the early releases will not be promising enough to attract new participants and the project will slow to a crawl or even stall out. I have no way of knowing which of these scenarios is more likely based on the data that I have now, but I will post updates on this topic in the coming months as the level of participation in Sakai 3 development becomes clearer.
Anyway, let’s see how many steps it takes to stretch an analogy ad absurdam:
Today’s Sakai 3: Mac OS X Public Beta
In the Fall of 2000, Apple released the public beta of the new operating system. It really had no value for end users, other than to give them a glimpse of the future. The $29.95 license fee provided them with little more than they could get from online screen shots and screencasts. But end users weren’t the target for this release. Developers were. The idea was to get developers educated and productive in the new system so that third-party applications would be ready soon after the OS was ready for normal humans to use.
Likewise, the code released for Sakai 3 to-date, which includes the new kernel (codename: K2) and the new user experience layer (codename: 3akai) are primarily aimed at getting developers productive and giving end users a glimpse of the future. Like Mac OS X but unlike the commercial LMSs, Sakai is heavily dependent on third-party application developers. There is no central Sakai development team to build the grade book or the discussion board. Those are applications that are contributed by independent development groups, usually led by one or more adopting universities. As a result, it was critical for the Sakai community to get code out that developers can play with well in advance of a product that is ready for end users.
Sakai 3 in 2009: Mac OS X 10.0
Mac OS X 10.0 was more of an opportunity for Mac users to get used to the new OS than it was a real production-worthy operating system. It was missing some features, it was slow, and it was just buggy enough to be annoying. It also didn’t have a lot of third-party software available. Theoretically, you could do generic computer work (there was a version of Microsoft Office, a working email client, and a working web browser), but the system performance was enough of an issue that most people didn’t use the system full-time. And if you needed to use specialized apps, you were out of luck. It was possible to run “Mac Classic” inside OS X, which meant you ran a virtual copy of OS 9 that could support most of the old Mac applications, but it was slow. (It also required a full installation of OS 9 on the machine alongside OS X.) In practice, most people booted into OS 9 most of the time.
There is some similarity here to the Sakai release that is due out at the end of the summer. Like OS X 10.0, it is intended primarily to give adoptees an opportunity to get used to the new system and to give developers a more complete system to work with. It will have a small handful of basic applications—discussions, polls, file sharing, blogs, and homework dropboxes—and some new features to the platform like social networking and drag-and-drop layout configuration. Cambridge University and Georgia Tech have committed to adopt this version for limited (mostly non-course) use, and I expect that other schools will install it and fool around with it. But for the most serious uses, most Sakai schools will be doing the equivalent of booting into Mac OS 9, i.e., using Sakai 2.x.
Sakai 3 in 2010: Mac OS X 10.1?
Mac OS X 10.1 was when OS X started to be a real usable option as a full-time system for at least some users. It was not trouble-free, to be sure, but particularly after the first point release, it was good enough to start booting into it by default. Gaps like lack of DVD playback and CD burning were fixed. Performance improved enough that you could do your basic work native in OS X and run legacy apps you needed in Classic mode in most cases. Late in in 10.1’s life, a native version of Photoshop was released. OS X began to feel like it was going to be more than just an interesting experiment.
We don’t know what the 2010 release of Sakai 3 will be like yet because we don’t know what kind of resourcing it will have, but it’s probably not realistic given what we know now to expect a leap forward that is larger than OS X saw from 10.0 to 10.1. There will likely be improved performance, and many of the tools that one needs for basic day-to-day LMS use will probably be there, but there will also probably be continuing gaps and rough edges. Two tools to watch are the grade book and the test engine. These are both very complex applications, and both of them run the danger of working awkwardly in the new world of Sakai 3 if they are ported rather than re-implemented. Will they make it into the 2010 Sakai 3 release? Will they be full re-implementations? If things go well, my guess is that the grade book could be re-implemented and the test engine ported. More than that would be a huge accomplishment for the Sakai community. Much less than that might be a sign of significant resourcing issues.
Like OS X, Sakai 3 is planned to have some sort of compatibility mode where a separately running copy of Sakai 2.x can be (awkwardly) integrated with a Sakai 3 environment. For current Sakai schools that rely on particular 2.x tools, this will be important. For potential new adoptees of Sakai, running the compatibility mode might not make sense, so it will be important for them to look at the state of Sakai 3-native tools and the roadmap for the 2011 release.
Sakai in 2011: Mac OS X 10.2?
For me at least, Mac OS X 10.2 was a watershed. To begin with, it was fast enough and fully featured enough that I didn’t miss Mac OS 9 anymore. All the apps that I needed as a general-purpose user were now native. I knew a few people who still had to use Classic for specialized apps, but I never had to touch it (or wanted to touch it) myself after 10.2 came out. It was also a watershed because we began to see applications for the Mac that you couldn’t get anywhere else. For me, the killer app was NetNewsWire, the first native desktop RSS reader on any platform. At that point, it became clear to me as an end user that the most powerful features of OS X, the ones that were going to impact my life the most, were the ones that I couldn’t see. They were the ones that made individual developers vastly more productive. (Ranchero Software, the company that produced NetNewsWire, was literally a mom and pop shop with one developer.)
If development goes well, then the 2011 release of Sakai 3 could be very similar. To begin with, the system should be good enough for most general-purpose use for most schools (and if it isn’t by this point, then the Sakai community will likely be in serious trouble). There will be a number of current adoptees who will still need to run compatibility mode for reasons of change management, but my guess is that this will be the release when major migration of the installed base begins (again, if all goes well). New adoptees shouldn’t have the same change management issues (although they’ll have different ones), and Sakai 3 should be very competitive at this point.
Most importantly, we should start seeing real innovation by the 2011 release. One of the major design goals for Sakai 3 is to massively lower the barrier to entry for developers. One person with HTML and Javascript skills should be able to do new and interesting things. Add in some skills with a web scripting language (anything that can consume RESTful services), and look out. Sakai 3 is also designed to support use cases that will make the anti-LMS crowd happy. It should be able to hide under the covers and provide just enough integration with tools out on the wider (and wilder) web.
Sakai 3 Beyond 2011: Much Coolness?
I have great hopes for Sakai 3. The vision is spot-on and the architecture is smart. A lot will come down to resourcing. Many of the core technologies in Sakai 3 are actually externally developed open source components (mostly Apache projects), which is critical. Without being able to rely on these larger communities to help support some of the harder stuff in the deep plumbing, Sakai 3 would be completely impractical. As it is, I think they have a real shot at pulling it off, but the community is going to have to step up to make it happen.
