WinFS, mutated or mutilated?

WinFS, mutated or mutilated?

  • Comments 3
 

I need to comment on last week's WinFS Update (and the Update to the Update), announcement by Quentin Clark, as the WinFS is something I've been very passionate about for the last couple of years. A couple of years ago, I delivered a WinFS talk on a roadshow (with Clemens Vasters, David Chappell, Nigel Watling, Lester Madden) to about 10.000 people. The enthusiasm was real, people could see how this was going to change computing in a good way for them as an end user and as a developer.

 

When I was at Tech Ed Boston two weeks ago, I attended Quentin's talk on the WinFS and I heard the story that was told ever since WinFS was ripped as out as one of the pillars of the WinFX. I too, was happy to hear that Beta 2 seemed to be on track. And then this; parts of the WinFS technology will move into ADO.NET and SQL Server but there won’t be a separate WinFS release anymore.

 

The comments have been mostly negative and I must admit that, to a certain extent, I do understand the disappointment. Originally, I believe, the FS stood for File System before it was changed to Future Store. The first naming expressed exactly what a lot of people came to expect; a file store that would make it easy for files to be searched, be shared (between devices and people) and be related to other things one would store. To me, that was WinFS and, of course, there was going to be a model to make developers productive at working all that data and that would some kind of ORM model.

 

There seems to be more clarity now about where the entity technology is going but less clarity on what the current vision is although Quentin does claim in the update to the update that the vision is still alive. This entire post is obviously my personal opinion but I definitely hope for a comeback of the full idea.

 

There are actually a couple more products Microsoft announced that I liked and that were dropped including: WinFS, MBF and .NET My Services. The one thing these technologies had in common was SCHEMA. Schema for common objects in the domain the technology was designed for. Objects in the personal computing space for WinFS, business objects for MBF and objects in the domain of people connecting to the web and to other people and devices for .NET My Services.

 

I definitely don't want it to come across as if the lack of agreement on the schemas - or the lack of industry support for the schemas - has sunk each of these projects but I think it's clear that they have been a big point of discussion. To really push forward computing, the industry will need to agree on a number of formats like file formats or formats to represent data on the web. It will require some giving and taking from Microsoft, other big players in the industry and the ABM-camp (anything but Microsoft).

 

Take for instance Microformats. They are a push in this direction to share schema on the web and although I'm certainly not going to judge on the quality or viability of this work, I certainly hope that one day we will have deep support on the web and the desktop for common formats of the most common objects in the most common domains. And if there's one company that can seamlessly integrate the web and the pc (including phones, media centers, game consoles, etc…), then it's Microsoft. And now you know why I work here.

This post is about: WinFS

 

  • *sigh* it’s sad that many Microsoft employees miss the point as to why the public is “disappointed”

    Your company is SO big that “politics” plays as big a role in developing your technology as writing the software. For the many Software Engineers that sat through seminars (as little as a 2 weeks ago) and studied how to use WinFS they feel (rightfully) deceived.  

    While Microsoft always offers logical explanations -- I’ve never seen them act “humble” and “apologize” for the “collateral damage” done to the people that actually use/support your software.
  • We still have a problem with the file system though.

    Look at the amount of files hold by every computer today, and look at the current NTFS, it can handle the volume, but with a very poor performance.

    In the other hand, Sun ZFS, and Linux are way ahead, and look at the source code, it is small and simple, and it does a lot!
    http://www.opensolaris.org/os/community/zfs/source/

    It is nice that one day in the future Microsoft will solve the data storage issues once forever, but for the next couple of years, it will be nice if Microsoft improves NTFS.

    All what is needed is a mini database like functionality on the kernel level, and that mini database used to store the basic file metadata (names, extensions, etc) the most basic elements, and keep the contents of the file for the application to manage.

    WinFS may handle all the possible combinations of all the possible ways to store data, that’s in the future, for now; I would just love NTFS2 if possible :-)
  • Hi Blake,
    I do understand and feel the disappointment. I even feel responsible for wasting thousands of people's time on this.
    When it comes to this type of stuff, evangelists are in the same seat as our early  adoption customers and partners.
    We are warned beforehand that everything we hear about is still beta and then we spend some/a lot of time on these technologies and sometimes waste some/a lot of times.
    The more we open up at Microsoft, the bigger the risk will be that some of that time will be wasted. I still think that being more transparent is good but we need to keep putting things in perspective.
    As to explaining why decisions are being taken, I agree that we still have a way to go.
Page 1 of 1 (3 items)