The footnotes (in parentheses) are extensive and should be reviewed with the full text of this article.
The successor to Windows XP (due in 2004, and rapidly slipping to 2005) is currently code named Longhorn, and it will not be compatible with your existing software, hardware or methods. Microsoft has already stated that backward compatibility will not be a design feature.
Some expect the name Windows will be dropped completely. The antitrust agreement with the Bush DoJ specifically states “Microsoft Windows” throughout. By maintaining incompatibility (already planned due to design considerations), making it look different and calling it something else, Microsoft can free itself from antitrust oversight. “It’s not Windows, it’s a different product – the agreement doesn’t apply.”
So this will be the “from the ground up” rewrite that XP was supposed to be, but not to remove bugs as we would hope.
The most important feature of Longhorn is replacement of the familiar DOS/Windows filesystem with an object database (W0). You will no longer copy files to a floppy or CD-ROM or attach them to an email, because there will be no files. Database records will be copied from one database to another, probably through a .NET server. Large organizations will have their own .NET servers, but everyone else will use one of Microsoft’s, a service for which you will pay a fee.
The Longhorn filesystem will be based on the technology of a re-thought and expanded SQL Server database (the project coded Yukon) (W8). Obviously, SQL Server being so tightly integrated with the filesystem (W19) will have a negative impact on publishers of other database engines for Windows. Not strange then that market leaders Oracle and IBM are heavily pushing the Linux platform and barely mention their products run on Windows any more.
Ponder that a moment. You won’t be able to manipulate files as files, and send them to someone or make a copy to take home. You will “own” some file handle or token that gives you access to the contents that you refer to as a “file” but that seems a little precarious.
Tim O’Reilly made a comment a year or two ago that stuck with me: in reference to the upsurge in interest in XML as a platform-neutral data storage medium, he expressed hope that “users could once again own their own data.” (I can’t find the original, but I think I have the sense of it in my paraphrase.)
Yukon seems to be the antithesis of that. And then, just a few days ago, Tim Bray was working on his Technology Predictor Success Matrix and listed object-oriented databases as a clear loser:
I remember being informed in superior tones, sometime around 1990, when I wanted to use boring old Sybase for some problem or another, that anyone who went to the big database conferences knew that the debate was over and that object-oriented had won.
Well, it wasn’t and it hadn’t. None of the pure OODBMS vendors are still alive, in that line of business anyhow; and if you proposed an OODBMS as part of a large mainstream application these days, you’d get some funny looks indeed.
Yes, conventional relational databases these days have some features gleaned from OODBMS field, but at the end of the day, this was a technology that just didn’t fly.
Well, one can only imagine what would happen if Tim Bray went to MSFT and how short a honeymoon he would have . . . .
If object-oriented databases are a non-starter, why is every Windows user going to be using one? And what’s the value of a database instead of a file system? Ease of searching? Speed of retrieval? Wouldn’t some kind of database that indexed the filesystem achieve the same ends?
My conspiracy/skullduggery detector gets jumpy when I read some of the stuff about Windows PC calling home to the mother ship . . . .
Gah. I’m just glad I’m not forced to use this junk anymore.