Re: X and other visions

[Alfred M. Szmidt]

   May I quote you?

That you have a crystal ball that can look into the future? Sure why
not.  But only if you make this crystal ball avaiable for everyone to
look into.

   You said once software wasn't something magic. ;)

And I still claim that it isn't; claming that it is shows that one
does not understand what software is and what it does.

   Software is made by human decisions, and what I told you was also
   told me by human persons.

Once again I urge you to read what I have written, software is not
magic and more or less claiming that you have a crystal ball that will
tell what people will use in the future are two completely different
things.

   >    (doesn't mean it won't support XFree86 any further, but I
   >    don't know ...)
   > 
   > X11 must be supported as long as X11 programs are used and as
   > long as the GNU desktop, GNOME, uses X11.

   XFree86 != X11

Never said that it was.  XFree86 is a implementation of X11, so
anything that runs on XFree86 will run on any other X11 implemenation.

   Have a look at XGGI.

Yes, lets,

 xggi-1.6.2.tar.gz       15-Oct-2000 20:52   26K  

I would call the project dead.  And since it is just a patch for
XFree86, then by your claim that XFree86 is a hack this makes xggi a
hack...

   >    You could also hack all the programs to use a user-specific
   >    logfile. Happy hacking.
   > 
   > Wrong, just fix syslogd (and/or syslog()) so that it can write to
   > a specific user-file.  All sane programs use this to report log
   > messages.  Then the log-monitor can watch that user-file.

   OK, happy hacking and recompiling.

Why do I need to recompile anything anyway? glibc implements syslog,
so I only need to recompile glibc; and all programs will magicly
support this wonderful feature.

As for hacking, your idea requires far more "hacking" then just
extending _one_ library call to dump logs into a seperate file based
on what UID a process is being run as.

   But I doubt end-user-space programs use syslog().

Anything (sane) that wishes to write to a log file uses syslog().

   > Yes, Paxillus, the music box, is not my computer.  You send a
   > song to it, it puts it on a stack, and pops it when the song is
   > to be played.  You can also log in on a port and set the volume
   > remotely.

   But wouldn't you also be able to change the volume when another one is
   playing a song?
   Would that be nice?

Yes it would, that you cannot see what it wouldn't be nice is not my
problem.  Take the scenario that someone cranked up lots of music on a
very loud volume setting, should I not be allowed to lower it?

   >    You may be annoyed if another one starts an X server on your
   >    machine ...
   > 
   > Why should I be first of all? And how is this different from starting
   > a ssh daemon running as a normal user on port >1024?  Or any kind of
   > daemon, which would include a X server into that set.  Are you going
   > to tell people what they are allowed to run and what they are not
   > allowed to run?

   That's usually an admin's task.

That it is usually the administrators task is irrelevant.

   What I'm talking about are only features, that may be disabled.

If you talk about features that can be _enabled_ (which they already
can), then this disucssion is moot since I won't object to such
things.

   You are the owner of /dev/dsp?

No.

   > Every user can become root, they can use a sub-hurd or even a
   > emulator.

   Wonderful.  Being root may be a pleasure, but getting permissions
   on the parent machine may be even better.

Huh? What are you talking about?

   > Huh? It is the "tty driver" that is providing this information;
   > just what you wanted.  And infact, you can do it simpler,
   > /dev/tty is always the currently controled tty by a interactive
   > processes.
   > 
   > In your shell you can type: tty, to see which tty is being used.

   Unfortunately the XFree86 hack is too bad too support it.

Why don't you back up these claims? `tty' for example shows perfectly
well which tty is being used by a xterm.  For "pure" X programs, you
can check that by checking which windows is being currently focused.

In ratpoison, my window manager of choice, what window is being shown
can be checked with the following key combo: A-q i (A-q is the escape
seqence which usually is C-t I belive).

   He is a troll because he only tells destructive stuff in a "Linux"
   forum.

That does not mean that the person is a troll if the things he say are
correct.  Or that you disagree with them.  Calling names is always bad
taste.

   "When you do basic file operations [I think he meant mv in this
   case], the ACL information of this files may be lost."

Because this probobly is correct, mv cannot preserve permissions over
all scenarios.  You can't do that if you are not the owner of the
file, or root.  And I might agree that this could be considered a
broken behaviour...

   > And that only root should be able to set a translator on a device
   > node that happens to be a real hard disk, is utterly stupid.  It
   > goes against ever sense of logic out there, if someone owns a
   > node, then that person should be allowed to do whatever they
   > please to it, _PERIOD_.

   I don't know what kind of data you store on your HD, but in some
   cases, it may be bad anybody is allowed to set translators to any
   not-mounted partitions.

I urge you a _third_ time to read what I wrote, I never said or
implied that _ANYONE_ should be able to do anything.  I said that the
_OWNER_ of the node should be allowed todo anything to said node.


Jeez...