Re: what's up with xterm?
OK, I had better clarify my statements.
Brandon Mitchell scripsit:
|Ok guys, lets go a little more easy on Branden. After all, he has done an
|excellent job with X, and he has a cool name :-) There is nothing wrong
|that I can see with the current setup of the $TERM setting. What we are
|asking is for an extra option during the install to allow the user to do
|the "wrong" thing. If you don't understand why we use xterm-debian,
|please read
| http://master.debian.org/~branden/xsf.html
I have read it, it is all very well but it doesn't fit with what I am
trying to say. Clearly I cannot express myself properly and I
apologise for that.
|On Fri, 20 Nov 1998, Arrigo Triulzi wrote:
|I didn't recommend that you do this. I said I would because it is the way
|I like things. I'm using a debian supported config and accessing remote
|systems without hassle. If you have never used CDE or other strange unix
|X windows setups, you should check the other message on this subject,
|which basically states that xterm-debian is a semi-norm to the degree that
|we aren't using xterm. I'm using CDE right now and know exactly what that
|poster is metioning.
OK, let me clarify. Systems over here adhere to a very simple rule:
X11 is the standard, plain X11. All workstations are either installed
with the optional non-CDE set (e.g. Digital UNIX), or have MIT X11R6.3
installed (older Suns, etc.). The window manager is fvwm and that is
that. Diversion from the norm is allowed at the user's responsibility
(i.e. no help). The reason for this is that we aim to have a uniform
environment (the lack of SGIs and HP-UX boxes due to excessive cost
helps :-)), and Debian has so far fitted splendidly.
People have been using the cheap PCs with Linux to access the
computation power of the workstations, using plain xterm. This is
why having xterm that works out of the box is rather useful.
|I object. Remote logins do work with a little hassle, and xterm-debian
|fixes some bugs that cause more of a hassle to the user than remote logins
|do, at least in my opinion.
Fair enough, that is your opinion, but how many foreign staff do you
have accessing machines around the globe running all sorts of UNIX
versions (ever tried Sinix? Quite popular in Germany on Siemens
machines)? Very, very few of these are UNIX literate beyond compiling
Fortran, running TeX and reading e-mail. In fact to them UNIX is a
different DOS or a different VMS. This also excludes occasional
visitors who of course get an account. The last thing you want or need
in a similar situation is to have to write a leaflet which half the
people won't understand explaining that they need to edit files they
have never heard before adding lines to change their terminal setting
automatically. Think about the number of shells, you have sh, bash,
csh, regional variations, quirks, I am sure you can imagine the
scenario. Or am I supposed to spend the day fixing people's remote
accounts?
|If you are giving people Debian linux to administer, but you don't want to
|explain what things like a .profile file is, you are asking for problems.
Well, it isn't that easy. You have lecturers who do not want to
release the root password to their machine, i.e. they want to be
responsible for the machine but of course if you are the person
responsible for running the UNIX boxes you have to answer their
questions. The less you generate the more you can concentrate on your
research and useful work. Unfortunately hierarchies are such that you
always have to take care of these people, whether or not they give you
the root password. The boss wants it this way, full stop. Also, you
cannot lecture lectureres on how to use UNIX because they all know
already :-(
|Explaining how to change settings for one areas setup is the norm, not the
|exception. Granted, this is one setting that would be nice not to change,
|but giving them no idea of how to configure a system for your local setup
|is bad idea.
There is no need. In the local setup all they need is IP address, DNS,
and a couple of trivial NFS mounts (/usr/local, /home, /var/spool/mail
if not a POP-3 user), that is all. They are given a beautiful list of
packages to install which they feed dpkg (copious thanks to Bruce
Parens for telling me how to do it back in 1.1 times) from a file. It
then installs a pretty machine which matches all the other staff
boxes and works just fine.
|This has already been considered. My understanding of it is that we can't
|change the settings for TERM=xterm, because all definitions of xterm on
|all setups of a unix-like system must be the same. So, if we want to fix
|things, we need an xterm-debian. And, if we want users to see these
|fixes, we need to make xterm-debian the default.
Fair enough. But it is still rather painful a change to implement? I
feel rather confused by the fact that I seem to be the only one
running a 100+ network of Debian boxes who is complaining or stumped
by the problem on debian-testing.
|> Please do consider not only making a system which "complies with the
|> Debian Backspace policy" but also which works with the rest of the
|> world. Or are we trying to copy the M$ "extend and embrace"
|> technology?
|
|Please avoid the low blows.
This wasn't meant to be a low blow but something which is in use in
the UK can is called irony, smilies were clearly needed. Apologies. It
just feels like we are adding an incompatibility which wasn't there to
cure a cosmetic problem.
If the powers that be have decided that the cosmetics is better than
the compatibility, well, OK, what can I do, I won't upgrade I guess
until a better solution appears (if ever).
Please understand that I respect deeply the work of the Debian
developers and contributors, mainly because it is a far better overall
system than RedHat for what I need, it is just that if I am a tester I
have to give a testing report. 2.1 as it is now (and 2.0 for that
matter, but I had no time to test it at the time) is *not* suitable
for running large networks of similar machines in a non-UNIX literate
location like mine. I know of other UK universities where similar
opinions are quietly suggested, perhaps we do things differently over
here.
Once again, it wasn't a low blow or meant to be, seriously.
Ciao,
Arrigo
--
Arrigo Triulzi <arrigo@ic.ac.uk>, http://www.ma.ic.ac.uk/~agbt
Mathematics Dept. Imperial College of Science & Technology - London - UK
--
To UNSUBSCRIBE, email to debian-testing-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: