Microsoft's Midori operating-system skunkworks project soldiers on

Microsoft's Midori operating-system skunkworks project soldiers on

Summary: A new research paper makes a not-so-thinly-veiled reference to Microsoft's secret project to develop a non-Windows-based operating system and programming environment.


Midori -- the Microsoft skunkworks operating-system project -- is still alive and moving forward.

Midori is a new operating system being developed by a team of all-star Microsoft programmers. Midori is not based on the current NT kernel; instead, its original roots can be traced back to Singularity, a Microsoft-Research-developed microkernel operating system. Headed (at least at one point) by Senior Vice President of Technical Strategy Eric Rudder, Midori is believed to be a distributed, concurrent operating system. The product and associated deliverables (a related programming language/framework, etc.) are in technical incubation.


The latest, not-so-thinly-veiled reference to Midori comes via a presentation at last month's OOPSLA 2012 conference. At that event, several Microsoft employees presented a paper entitled "Uniqueness and Reference Immutability for Safe Parallelism." The paper outlines a prototype extension to C# that extends C# so that it supports safe task and data parallelism.

From that paper:

"Our type system models a prototype extension to C# that is in active use by a Microsoft team. We describe their experiences building large systems with this extension....

"A source-level variant of this system, as an extension to C#, is in use by a large project at Microsoft, as their primary programming language."

This "large project" is the Midori project.

A Microsoft job description for a software architect to help "lead development of our safe concurrent programming model" for the Midori team makes reference to some of the same work outlined in the OOPSLA paper:

"This programming model is a core component of a new, novel operating system, 99% of which is written in type- and memory-safe C#. A core principle we add to managed code is that 1st class, statically enforced concurrency-safety must become a peer of type- and memory-safety. This role demands innovation at each layer of the software stack: programming model abstractions, scheduling (kernel and user-mode/runtime), message passing and asynchrony more generally, shared-memory, data and task parallelism, distributed parallelism, heterogeneity (including vectorization and GPGPU), interaction with processor architecture, feedback directed optimizations, and even language design and compiler implementation."

I found the link to the OOPSLA paper (thanks to a tipster who asked not to be named) via posts on both Rob Jellinghaus' and Joe Duffy's blogs. Both Jellinghaus and Duffy are known to be working on Midori.

Speaking of who is/isn't working on Midori these days, thanks to some sleuthing by "Felix9" of Microsoft Channel 9 fame, we know that former Midori team member Rick Ledoux has left the effort and is now working at Buzzhorn. Matthew Moore, a Softie who has been working on the "Phoenix" compiler framework developed by Microsoft Research and the Microsoft product compiler teams, left Microsoft recently to join Google. (Phoenix has been touted as "the universal compiler backend for upcoming Microsoft languages and development tools.")

Meanwhile, another name to add to the current Midori team roster is Shon Katzenberger, a partner software architect at Microsoft who has worked on another Microsoft incubation project (Tangram), as well as Microsoft Solver Foundation, C# compiler and language design.

Not every Microsoft incubation project ends up seeing the light of day. But given how long the Softies have been working on Midori, I'm thinking this one will end up going commercial at some point.

Topics: Software Development, Microsoft


Mary Jo has covered the tech industry for 30 years for a variety of publications and Web sites, and is a frequent guest on radio, TV and podcasts, speaking about all things Microsoft-related. She is the author of Microsoft 2.0: How Microsoft plans to stay relevant in the post-Gates era (John Wiley & Sons, 2008).

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • So when are we gonna see Microsoft Word for Midori?

    They have been working on this thing for a long time, makes you wonder if it will be forever a research project.
    • Too busy

      MS is up to its a$$ in alligators in mobile. I cannot see them being distracted by some totally new OS venture. That could be the end of the company.
      • really

        Are you on smack?

        You do understand just how much money MS sinks into R&D dont you?
      • it seems you have a lot of information on Microsoft's internals

        Could you please care to share before you shove your "golden words" here. AFAIK, Microsoft spends more $ on R&D than your lover Google.
        Ram U
    • Research

      Most research projects like this are never intended to be productized. The typical use is that bits or ideas will get folded into existing projects. The problem (for us) is that when that happens we rarely know it -- it just looks like normal product maturation.
      • Midori is not a research project

        Hi, just so you folks know: MS is not categorizing Midori as a MSR project. It is a technical incubation, which means it is far more likely to end up commercialized at some point. MJ
        Mary Jo Foley
        • Awesome!

          I've been following this project casually since I first heard about it a few years ago. Personally, I'm VERY excited to see where this goes. I don't care who is behind it, the tech is what I'm interested in. Every now and again *someone* does something so groundbreaking it changes everything. I think a project like this has that kind of potential.
        • Windows 9 or 10?

          Which means it is in line for possible productization.

          Perhaps by Windows 9, very likely for Windows 10.
          Ian Easson
        • Indeed, Looks Like MS Still Hiring for Another Midori Position

          See The exact title of the position listing? "Software Developer engineer II for Midori Group" . So if you'd really be thrilled to work on the Midori project, are qualified to do so, and are willing/able to move to Israel (or, more likely, are from or already live there), well, there's a link to start the application process on the page.
  • Have any bits...

    ...been used from this project in currently commercial projects? The Midori projects seems like one of those projects that over time can develop value components that can be utilized in other existing products to improve them.

    Very interesting.
  • hot air!

    [i]This programming model is a core component of a new, novel operating system, 99% of which is written in type- and memory-safe C#. [/i]
    That's doomed from the start.
    IBM tried with a Java OS in the 1990's and failed. M$ has no chance of doing it better.
    LlNUX Geek
    • really

      the biggest most succesful software company ever who spend more in R&D that pretty much every company in the world cant do better than IBM?
      • Good point.

        They can dump billions at it. After all, look how much they spend to advertise Internet Explorer (!) and the stockholders don't seem to bothered by it.
    • Wow...

      In all of the political banter on some of the mainstream sites I thought I saw the pinnacle of ignorance. Are you serious? Someone else tried it and failed (20 years ago), so it's not possible. Everything that hasn't been done before is impossible. Got it...
    • It's not doomed...

      ...if they make the C# source compile to native code using some kind of verifying compiler, for example. Singularity was a glimpse on what such an environment could look like, and it's definitly very interesting. It's way superior to what IBM did with Java OS back then - we're living in 2012, it's not the early '90s anymore.

      I think it's important the re-think the way operating systems should be designed today - it would be very silly to hold on to patterns and rules that were created 30 years ago. As a student, you could still start and play with the Minix kernel, but with all the progress in technology, how much sense does that make today?
  • You can fill a bookshelf with OS Code

    So I wouldn't count it out as a commercial product one day. One would hope that building an OS from the ground up would take a very long time.
  • And to think...

    ...I was railroaded out of a job of 24 years because, among other "reasons," I resisted learning I asked them why, if I'm learning a programming scenario so different from the LAMP I was used to, shouldn't I at least be future-thinking about it, and learn C#?

    ***NO*** was the thunderous reply from the VB dinosaurs...
  • Will It Still Have Drive Letters?

    Knowing Microsoft, it will.
    • Origin of Drive Letters

      Drive letters are not the purview of Microsoft. They actually come from CP/M, developed by Intel.
      Even your printer had a letter designation! You would simply PIP (copy) your file to, say, E: to print it.
      Microsoft simply happens to have licensed the OS.
    • What's So Bad About Drive Letters?

      Especially for uses that are not tech freaks, the concept of drive letters is easier to understand than mount points in a Linux file system... they're just some kind of shortcut of the latter. Under the hood, the NT object system works pretty similiar.