Re: Of mice and men


In article <ma3fe.31677$3U.1402290@xxxxxxxxxxxxxxxxxxxxx>, Donald Tees <donald_tees@xxxxxxxxxxxx> writes:
> jce wrote:
> > "Donald Tees" <donald_tees@xxxxxxxxxxxx> wrote in message
>
> >>Directories and files under Linux are *owned* by a specific user ID
> >
> > You can also do this in Windows and others
>
> Other's yes. Windows, no.

All filesystem objects (for real filesystems) in every current
general-purpose (not embedded) version of Windows have ownership and
access control. This claim is simply wrong.

(That Windows continues to support the primitive FAT filesystem is
irrelevant; so do most versions of Unix, and they can't do anything
about its lack of security either.)

> >>You might be able to CD to another user's account (IF and ONLY IF you have
> >>access access), but that does *not* mean you can write to it, nor does it
> >>mean you can read it. You must, for someone else to even see the
> >>directory, give others that access.
> >
> > You can also do this in Windows and others
>
> You keep saying this. Are you talking about NT server? That is not
> windows.

As is this one. NT Server (and NT Workstation) is most certainly
Windows; I don't see any basis whatsoever for pretending it isn't.
And filesystem permissions are present in all the versions of
Windows that are derived from NT, which includes *all* versions of
XP. That applies to the current "home user" version of Windows,
XP Home Edition.

> > Funny that my college account was windows based and I didn't have access to
> > anything then....I wonder why Microsoft created Users/Groups and Access
> > Control Lists and cluster servers if they didn't have any ownership
> > protection....
>
> Because the Unix servers they were attaching to did. They predated NT
> server software by several years. In fact, users's and groups were
> available in DOS as net user commands.

The NT team did not include user accounts and DACs because of Unix;
they included them because they were a minicomputer-OS development
group. Cutler came from VMS.

When NT was developed very few Unix systems were running filesystems
that supported ACLs. There were some, such as AFS, but they were a
small minority.

The DOS "net user" commands set user credentials for SMB connections
only. They weren't part of DOS any more than entering a username and
password in a Telnet session from a DOS client were.

> The only thing they have ever been used for in windows/DOS is to join a
> network.

That is completely, utterly incorrect.

> That is why the cancel button logs you in ... you are not
> canceling login to the computer, there is none, and there is no
> security.

None of those versions of Windows are current. They're unrelated to
the current OS except in name and some look-and-feel aspects.

--
Michael Wojcik michael.wojcik@xxxxxxxxxxxxxx

Reversible CA's are -automorphisms- on shift spaces. It is a notorious
fact in symbolic dynamics that describing such things on a shift of finite
type are -fiendishly- difficult. -- Chris Hillman
.

Quantcast