Tag Archives | Midori

Who Needs Syncing?

Nearly two years ago, I reviewed internal documents about Microsoft’s plans to design and develop an entirely new operating system called Midori. While I am uncertain about the exact state of the project, bits and pieces of the Midori vision are emerging in the company’s latest technologies.

Owning a PC was once a big deal; now it’s common for multiple computers to reside under one roof. Today’s households are filled with PCs, Pads, and Pods–devices that are loosely synchronized and loaded with apps. Information and applications are getting distributed, with many pieces working in parallel. Midori is intended to support exactly that kind of distributed application architecture, and Microsoft assigned some of its top talent to support the project.

Continue Reading →


“Barrelfish”: Microsoft’s Latest Future OS Project

Last Friday, Network World reported that Microsoft’s research labs in Cambridge (UK) has previewed an experimental operating system code-named ‘Barrelfish.’ However, it is just one of many fish in Microsoft’s barrel, and is not nearly as close as Microsoft’s project “Midori” is to becoming an actual product.

Barrelfish and Midori tackle a similar problem that Microsoft has determined cannot be met by evolving its existing technology. They run on multi-core systems, and are designed for heterogeneous hardware environments, where applications and resources can exist in separate places.

Beyond sharing a similar mission, there are major technical differences between the projects. Midori is rooted in a research project called Singularity, which is constructed using Microsoft’s .NET Framework; Barrelfish uses some open source components.

Barrelfish, like Singularity, is just a research project. Midori is differentiated, because it an offshoot from the Singularity lab work. Microsoft has placed Midori under the control of Eric Rudder, senior vice president for technical strategy at Microsoft and an alumnus of Bill Gates’ technical staff.

The company has also mapped out a migration path away from Windows to Midori, but there was still a lot of hand-waving in the memos that I reviewed last year. Microsoft has since placed all information regarding Midori under lock and key on a “need to know” basis.

After I wrote my Midori expose, I was told by a source at Microsoft that I had just scratched the surface. Microsoft is a big company that has a lot of resources, and I will not pretend to know everything that is going on in its skunkworks. What I do know is that Barrelfish is just research–for now.

No comments

Google Ties Chrome to Cloud Services

chromelogo5Today, Google fired a new salvo in the browser wars, announcing an upcoming synchronization service for its Chrome browser. A preliminary mockup of the service will be released to developers later this week, with general availability possible later this month, according to reports.

The service will first deliver bookmark synchronization –something that’s already possible with Firefox via plug-ins as well as Opera. Google will add other types of browser data incrementally. If Google carries out its plans effectively, Chrome will provide users with a seamless user experience across many devices. Other browser makers will have to follow.

Netbooks, which have the focus of Google’s most ambitious development efforts, will be an obvious beneficiary. The synchronization service will also give a boost to OpenID, which Google users to authenticate digital identities (with its own proprietary twist).

All in all, Google is continuing to blur the line between desktop software and the cloud. It is not alone in its thinking–I’m convinced that Microsoft, which is often perceived as its biggest competitor, will eventually follow suit.

Last year, I detailed Microsoft’s Midori operating system development plans. While Google has not announced anything as ambitious as Midori, it is going down the path that Microsoft laid out in the memos that I reviewed.

One of Microsoft’s principal  design motivations is to support the ability of users to share resources remotely, and for applications that are a composite of local and remote components and services. The Web browser is just beginning to enable the application side of that vision.


Microsoft: Windows CE is Closer to End-of-Life than Windows

On Monday, Microsoft debuted its long-awaited Windows Mobile 6.5 update at the Mobile World Congress in Barcelona. But as Microsoft moves forward with its “Midori” operating system incubation project, the longevity of Windows CE–the platform that Windows Mobile is based on–is in doubt, according to internal company documents viewed by Technologizer.

The documents refer to a “general sense that Windows CE is far, far closer to its end-of-life than Windows.” Indeed, Microsoft does not believe that the existing ecosystem of Windows CE devices, applications, third-party developers, and customers will force it to continue developing Windows CE indefinitely. “The device space is evolving fast enough that legacy support is less of a concern (than it would be to migrate users from the existing Windows code base,” it noted.

Midori is a componentized, Internet-centric operating system being architected from the ground up by a team led by Eric Rudder, senior vice president for technical strategy at Microsoft. Last summer, I reported that Microsoft was considering creating a layered, thin platform for mobile devices out of it–a product which could replace Windows Mobile over the long haul.

In such a scenario, Microsoft might end up with variants of one basic OS platform–Midori–on both traditional PCs and other mobile devices. The caveat is that Microsoft remains uncertain how far the company can go with a single codebase at Midori’s core.

According to the internal Microsoft documents, “There’s a limit on the hardware range that can be addressed with a single codebase, no matter how factored and substitutable the components. The litmus test of whether a device is too small to support with code may be if it can download code. Even if we cannot execute our code on tiny devices (light switches & smart sand), we may be able to extend our model to those devices, in the form of naming, the protocol for remote calls, etc.”

The documents noted that a general-purpose device OS must be carefully designed to be reconfigurable into myriad of configurations. That is not impossible to accomplish–Apple did it successfully with OS X, which powers both Macs and iPhones. Microsoft’s Visual Studio development tools are being updated with Application Lifecycle Management that will make it easier for it to create multiple versions of Windows based on the same codebase (assuming that Microsoft uses Visual Studio to develop Midori).

As I previously disclosed, Midori is designed to have a single framework for all device types called Resource Management Infrastructure (RMI). RMI is designed to manage and monitor I/O bandwidth, memory, power, and other resources, and to take them into account as it schedules tasks for processing. Microsoft believes that Midori’s power-based scheduling would be a good fit for mobile devices.

“I think we’re about to deluged over the next few years with Mobile Internet Devices as Intel pushes the Atom chip and AMD rushes to catch up,” Forrester Research principal analyst Jeffrey Hammond told me. That would seem an interesting place for Midori as the hardware is pretty similar [to traditional PC design], but power and form factor are different.’


Microsoft’s Post-Windows “Midori”–It Must Exist

It’s always dangerous to get too excited about far-off Microsoft products with code names–especially when Microsoft has barely acknowledged they exist. But SD Times has published a story by David Worthington on “Midori,” which Worthington says is the operating system that Microsoft is building from scratch for the post-Windows world, and even if you read it with a very skeptical eye, it’s a significant piece.

Worthington’s piece is at a site aimed at developers, is pretty technical, and is about pie-in-the-sky goals rather than specific features. I’ll do my best to summarize, interpret, and translate into plain English:

–Microsoft says that Midori is a research project; nobody knows when it might become a shipping one;

–Microsoft is working on a migration path to help folks move from Windows to Midori, as well as some means of running Windows apps in the Midori environment;

–Rather than being a desktop operating system, Midori will be distributed–that is, it’ll consist a bunch of components that can run locally on a traditional PC or remotely across the Net, and which can access data here, there, and everywhere;

–It’ll be designed to be more reliable, with a better understanding of what applications are doing and greater ability to prevent misbehaving apps from causing trouble;

–It will also put more constraints on software developers designed to prevent them from writing problematic applications in the first place;

–It’ll be designed for a world in which multi-core CPUs and other technologies enable massively parallel processing–computing jobs getting divvied up into smaller jobs that all happen at the same time;

–it will be designed to run on PC hardware or in virtualized environments;

–it will have sophisticated means of managing tasks and processes, some of which will relate to doing things in a power-efficient way, thereby making it an attractive mobile OS.

Again, Microsoft hasn’t confirmed the details of Worthington’s piece, and some of my interpretation may be off; the above items are likely more possibilities than confirmed details, and could be just plain wrong. Don’t start lining up at Best Buy just yet–and even if Worthington has his facts right, be prepared for Midori to evolve into something radically different, or to die the ignominious death that many intriguing-sounding Microsoft research projects have died.

Robert Scoble, whose opinion I respect, says that anyone who thinks that Microsoft will have a brand-new OS ready in the next few decades is–Robert’s word–an idiot. (He puts it another way: Bill Gates won’t be with us by the time an all-new Microsoft OS debuts.) I’m not so sure about that.

I don’t know how solid the details of Worthington’s report is. But I’ve got to believe that the broad strokes are correct, and that Microsoft is working on something which it hopes to turn into a product in years, not decades. It’s so utterly clear that the Internet is the computing platform of the future and that basic aspects of Windows are profoundly archaic that Microsoft would be crazy if it didn’t have people starting with a blank slate and working on figuring out what’s next. And the world is changing around Microsoft at such a fast clip that it doesn’t have decades to get its act together. The scenario Midori describes is gonna happen, whether it’s Microsoft, Google, the open-source community, or some company that doesn’t exist yet that makes it happen.

Of course, “brand-new” is a tricky thing to define. I don’t know if there’s any actual code from MS-DOS in Windows Vista–in theory there shouldn’t be, since Vista descends from Windows NT, which was allegedly the first version of Windows written from scratch rather than bolted onto DOS. But there’s no question that Vista carries a fair amount of legacy that dates back to DOS. In the 27 years that Microsoft has dominated operating systems, there have been no true big bangs.

So perhaps Midori, in whatever form it does take, will owe a lot more to today’s Windows than Worthington suggests. Every time Microsoft has said it was working on a radical shift in the past, the end product has proven to be less than radical; if Midori follows that pattern, bits and pieces of it will end up being integrated into an OS that’s evolutionary, not revolutionary.

But to think that Midori or something like it doesn’t exist is to believe that Microsoft is unimaginably dense and complacent…