Re: NTFS-like streams?

From: Kai Henningsen (kaih@khms.westfalen.de)
Date: Mon Aug 14 2000 - 15:15:00 EST


viro@math.psu.edu (Alexander Viro) wrote on 14.08.00 in <Pine.GSO.4.10.10008132344340.194-100000@weyl.math.psu.edu>:

> On Sun, 13 Aug 2000, Jeff V. Merkey wrote:
>
> > > Sigh... And then somebody sets "fascist" (read: reasonable) permissions
> > > and that program breaks. Right?
> >
> > On Windows 2000, you can create an attribute with any name and as many
> > of them as you want. When you open the file, you access the attribute
>
> "You"? What does it mean on multi-user OS?

Well, what does it *usually* mean?

Just apply the usual damn access controls. You don't actually think NTFS
doesn't do that, too, do you? (Or networking under MacOS.)

It's nothing particularly new. *This* part of multi-user those other
systems have already covered.

Hell, think of what happens when you pit in a CD. How do you change
attributes on a CD? Simple: you don't, and nobody is all that surprised
about it.

Don't create problem out of thin air.

> Trying to force the forks into the OS built on fundamentally
> different ideas is just plain silly.

Which OS would that be? Not Linux, in any case. NT (or MacOS) copy *so*
many of the basic Unix ideas (for example, the hierarchical filesystem, or
the idea of file-is-a-stream) that this claim just won't wash.

> BTW, the desktop crowd would do much better if they would start
> with making the X objects hierarchy visible as a userland fs.

Userland fs support for the kernel would be *very* useful for quite a
number of different applications.

>That would
> allow much simpler integration of CLI stuff and normal programs in
> general. Filesystem is The API on UNIX and all UNIX userland is built to
> work with it. Exporting your functionality via that API rather than
> inventing incompatible kludery would be really nice thing. Sigh...

And I bet that was part of the motivation of having forks in the first
place. "We need this additional stuff, we already understand files ..."

MfG Kai

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Aug 15 2000 - 21:00:35 EST