| d2r diego's weblog |
Cairo (and WinFS) revisitedInterestingly enough Jon Udell, whom I mentioned in my post about Cairo/Longhorn the other day, was writing up a great article with similar ideas for InfoWorld, including his own take on Cairo, Longhorn, and what the web is and how MS can affect it: [...] the Web is much more than the browser. It's an ecosystem whose social and information structures co-evolve. Innovation bubbles up from the grassroots; integration can happen spontaneously; relationships cross borders. Cairo Version 1 wasn't designed to nourish that ecosystem or to flourish in it. Let's hope Microsoft remembers the past and avoids being condemned to repeat it with Cairo Version 2.Agreed 100%. In an entry on his blog he points to my own post, and talks about the history of how he took it on himself to maintain Byte's archives, and some of the problems of maintaining "digital continuity." (I've been thinking about that too--more on soon). Then, in comments, Ole pointed to X1 which does full text search on Windows PCs as a similar idea to WinFS, Longhorn's new filesystem et. al. I have tried X1 and found it interesting, but it freaked me out by showing up all the time and being surprisingly difficult to remove. So I stopped using it. Others might have a better experience though :) Finally, Robert replied to my post (mis-spelling my name once again--by now this has all the makings of an in-joke ;-)), saying that WinFS vanished because I've heard that this is the fifth time that we've tried something like WinFS, but previous tries never got out of the lab. Why not?Hm. This explanation is not complete. Suppose you show it to users. They hate it. What do you do then? Ditch it, or go back to the drawing board? Well, it depends (as I said on my post) on resources. With a company of MS's size and resources, the only thing that explains ditching WinFS is a change of priorities. Otherwise you keep at it until it works. (And, again, I refuse to entertain the notion that "it couldn't be done"--especially when others have done it!) I'll stick with this idea until I am convinced otherwise. :) This still doesn't explain what happened to all that code either, or what, exactly, didn't work back then. Was it the speed? (Ole, btw, has been asking about that issue for a while now, and hasn't seen a good reply for it AFAIK). And that aside, the fifth time?? Really??? If that is true, it's a scary thought. The technology is definitely doable. I certainly can't believe (as I said) that Microsoft can't do it, so it sounds like there's a lot of resource-shifting going on. Weird. Categories: technologyPosted by diego on November 23 2003 at 4:56 PM Comments (please see the comments & trackback policy).
You've missed a point about making it work that's rather important - it not only has to work it also has to be useable and in a way that is accessible to Microsoft's users, all of them... Now as Linux demonstrates all too well, whilst "making it work" is easy making it useable to users who don't share our appreciation of the technology is significantly more challenging. Murph Posted by: murph at November 23, 2003 7:21 PMThis is true. However, the example I had in mind is not Linux but Mac OS. I still think that this is an issue of resources and not complexity. Besides, a lot of what WinFS is supposed to be will be hidden from users. It's hard to see how file metadata and better search that is supposed to be managed automatically anyway (and so is of most value to Office, ISVs, etc) makes it "hard for users". Posted by: Diego at November 23, 2003 7:44 PMIt's probably been approached by various different divisions/departments. The SQL people probably tried and the Office people (each for their own reasons) and then eventually someone must have got to the OS manager who decided it would be great to get it built into the OS itself. At which point it was considered worthwhile. It has to be said, having a database built into the OS itself would be darn handy. Posted by: Andrew Ducker at November 23, 2003 10:34 PMHi, Excellent piece as always Diego. I tossed up my pet theory here: Scott Posted by: Scott Johnson at November 24, 2003 11:54 AMRe:speed - BeOS was functionally equivalent to WinFS from the user's point of view; on what are now antique systems (300Mhz, wimpy video card, little memory, etc.) there was no discernible difference between a saved search and a regular folder. I managed email using Postmaster which had query-based virtual folders and again there wasn't a perceptible difference even with thousands of messages on a disk with 100,000 files. Considering that if Microsoft delivers they'll have a system as capable as BeOS (and, if we're really lucky, anywhere close to as responsive even with something like 6 generations of Moore's law helping out) almost a decade later, it's pretty obvious that it's not a resource problem. (A similar comparison could be made on the advanced compositing system with Apple, which is neither as well funded or well managed) Also, it's a bit snide but . . . not shipping a product with a bad UI or poor performance must be a very recent development at Microsoft - I just hope it doesn't start and end with Longhorn. Posted by: Chris Adams at November 25, 2003 4:54 AMCopyright © Diego Doval 2002-2007.
|