Live Mesh as the next information bus
Steve Gillmor sees a bright future for Microsoft’s latest initiative: Live Mesh. I do too, however, I have yet to try out the technology preview.
For now I’m content (well, pretty much–I’m eager to try out the preview, I’m just too busy to do so) with reading about what the Live Mesh team is doing, how they see Mesh being used, and where it might be headed.
As I do so, I pay particular attention to the DNA. When trying to speculate about where this or that technology goes, it’s often quite important to pay attention to where the foundation lies for the technology. It often sets the direction or bounds for its particular implementation. In this case, it looks like Live Mesh is rooted in Groove and Sharepoint think with a sprinkling of file sharing thrown in. That’s sounds a bit too corporate for me, but then again as Ray Ozzie pointed out in a recent Channel 9 video, part of the goal with Mesh is to bring some corporate distribution technologies to the mass market. The more I hear this–particularly coming from anyone at Microsoft–the more I become concerned that there’s going to be bounds to this technology spreading. Going corporate think to consumer think can be tricky. Going the other direction is much easier.
Awhile back I worked in a startup that developed a distributed application that sat ontop of the Tibco bus. (The application was designed to push out commands and updates to hardware devices in a network.) It was a powerful model. I see many simularities between it and the potentials of Live Mesh.
Steve Gillmor talks about the potential value of the Mesh-Silverlight pairing. No doubt that this could be a good combination since both are network biased. However, I hope the Mesh team is not thinking about coupling the two. A mesh network should not dictate this or that user interface. Silverlight might make it easy to connect to the mesh, but the two are distinct. From what I’ve heard from Microsoft, it appears that they are going in the right direction here and not tightly coupling the two.
I also wonder: when might we see a mesh language? When working on the Tibco-ish app I mentioned earlier, we ran into this issue. You have this information bus and all these apps and now how do you easily control everything? Our solution was to develop a language that was core to the Mesh itself–something very declarative. I’m wondering if the Mesh team is thinking along these lines too. It’ll be interesting to see–especially with the growing acceptance of WPF style development. This may reveal more about my age and when I started programming than anything else though.
It’s interesting to read some of the reactions to Mesh over on Robert Scoble’s blog. Sounds like people are having trouble getting passed the Microsoft-created and file sharing issues. In terms of the Microsoft-side of the equation, I see this as a plus. Any mesh technology is going to need wide distribution. Theoretically you can get this from the web, but Microsoft can support it more directly in large numbers. This is a good thing for the technology’s potential adoption. Also as a company, they are keenly aware of security issues, since they are such a large target. Again, I see this as a good thing for any new technology. Now in terms of the file-sharing aspects, they may have a point. There are huge potential upsides here, but how far will Microsoft actually take the mesh concept? We’ll have to see. I guess it all depends on how much of an early adopter one is. I know two people that have already downloaded and tried out the Live Mesh preview. Both grew perplexed and set it aside. That’s OK, we’re all trying to figure out what this Mesh thing means.