Linux-Setup Digest #495, Volume #19 Mon, 28 Aug 00 03:13:08 EDT
Contents:
Re: Linux, XML, and assalting Windows ("paul snow")
Re: Linux, XML, and assalting Windows ("paul snow")
Re: how to block Ctrl+Alt+Del ("David ..")
Re: Linux, XML, and assalting Windows ("paul snow")
Re: Belgian keyboard layout in Suse 6.4 (Geert Haustraete)
Re: Linux, XML, and assalting Windows
Re: Linux, XML, and assalting Windows (Darren Winsper)
Re: changing X-window appearance (Eric Carlson)
Recommendations on a good book. ("Adam H.")
Re: X windows (Daryl Fonseca-Holt)
Re: debian potato - selecting 'profile' to install now gone? (Michael Perry)
----------------------------------------------------------------------------
From: "paul snow" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.text.xml,comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux, XML, and assalting Windows
Date: Mon, 28 Aug 2000 05:26:10 GMT
<[EMAIL PROTECTED]> wrote in message
news:8o9s07$c3b$[EMAIL PROTECTED]...
>
> paul snow <[EMAIL PROTECTED]> wrote in message
> news:mZTp5.18778$[EMAIL PROTECTED]...
>
> > Today we have Java VMs and Adobe Acrobat viewers, and browsers, and
> browser
> > plug-ins, word processing packages, and stock tickers, Internet based
> games,
> > etc. Never mind that we are going to be configuring systems to connect
> with
> > other systems, and use databases, and database clients, and we need to
set
> > up security, and down load the new versions of our clients, etc.
> >
> > It isn't going to be simple in the future. It is going to get worse.
In
> > another post I list a set of requirements we are going to need from a
> > package manager. Not want, need. Typewriters are out for good. And
> > Redmond may be at fault to some degree, but if so they only pushed us
> ahead
> > in time a bit. It was going to happen to us anyway.
> >
> > We have to have package managers, but they need to be based on open
> > standards.
>
> Are you saying that the Linux package managers are not open?
No. But I don't have an open standard for describing software so that I can
express that software onto any platform its description claims it supports.
> > And they need to operate in an environment outside the execution
> > environment of the supported computer systems. They need to be able to
> > manage cross platform, distributed applications.
> Just what do you mean by that? It sound like a lot of empty talk devoid
of
> real meaning.
Distributed applications mean that to get the distributed application to
run, I have to get a number of applications installed and configured,
usually across a range of platforms, and integrated with firewalls, etc.
Today I do that with sometimes great programs (like the RPM) and sometimes
with less great facilities (like Install Shield). But I have to manually
install and configure each component, often in an exact order. I have
databases, database clients, compilers, web servers, object request brokers
of various sorts, etc. that all have their own sets of needed tweaks based
on the topology we are setting up.
All of that just to get the right abstractions up and running so we can
begin to talk about setting up and configuring the actual application.
That is what I mean when I say we need to be able to manage cross platform
distributed applications. Today we have given up on client side
distributed applications. Involving the client in all of this seems like it
is just too difficult. But look at the promises! Many of the applications
we are predicting require client side intelligence. And even the ones that
don't are passing up on lots of CPU cycles that would greatly improve their
performance. All because we can't deliver and manage software.
>
> > Why? Because we are on
> > the Internet already!
>
> What does that have to do with software installation procedures?
The Internet means distributed applications. And internet delivered
software. Transient software (applications that install, you use for a
while, then they go away.)
Take one of many examples:
Say that I want to be able, as an Application Service Provider (ASP), to
deliver native code to my clients. But today I can't because I can't manage
all the combinations of possible configuration issues that my customers are
going to encounter if I deliver software the way we do today. As an ASP I
become responsible. If I try to push this responsibility off onto the
specific Application Developer, they are simply going blame the problem on
one or more of the other venders that delivered software to this customer.
But if instead I am providing XML described software, then this
responsibility to deal with these combinations shifts to the vender of the
Software Rendering Facility the customer is using. But unlike the ASP, the
Facility vender deals with this day in and day out for all their customers
(a set much larger than the ASPs). And doing so is part of their value add.
Furthermore, part of their value add is assuring the customer they can
cleanly remove the new software if no repair for the installation is forth
comming.
That is why the Internet changes things.
> > We want to bank, we want to order hamburgers on the
> > Interstate Hwy so I don't wait for my order! I want to use my PDA to
> adjust
> > my lights in my hotel (cause I don't know where the switches are, but I
> have
> > my PDA), I want to listen to my MP3 files on the rental car's stereo,
from
> > the station I programmed on the Internet.
>
> Speak of your desires and do not attribute them to others. Still I ask,
> what does that have to do with software installation procedures?
We are at the breaking edge when it comes to delivering and deploying these
technologies. System upgrades? Replacing a busted server? Reconfiguring to
use the next release of a database? These are all very, very difficult when
they must be done manually, system by system.
What is to be done if even reasonably simple applications that you happen to
desire (but I would never dare imply intersect my own) happen to require
code to be installed on your computing device? Even if we "trust" each
other, my code might not be compatible with other code you are using. This
problem may even be a known one, along with the configuration fix needed.
How do we upgrade and manage distributed applications when they do break out
of the servers and onto the Internet in peer to peer configurations?
>
> By the way, do you know what it means to program a station? I don't think
> you do, or else you would not have used, "the station I programmed on the
> Internet", since it does not belong in this discussion.
www.Sonicnet.com I am not sure why I referenced it, but I just "programmed"
my station, which delivers western music, classical, and pop. You really
haven't lived until your radio station follows Bach with Johnny Cash! No
way anyone could have followed that reference, so I am sorry for any
heartburn it caused.
>
> > We can't do all of this by coping all our files onto our bin directory.
>
> I never mentioned anythng about copying anything into bin. I never even
> mentioned bin. Further, I would never suggest putting ALL our files into
> bin let alone onto bin. The installation method has nothing to do with
what
> the software can do once it is installed.
Forget the reference to bin (a really dirt simple install method that likely
predates you a bit).
However, you nailed it on the head there! Let me repeat you:
> The installation method has nothing to do with what
> the software can do once it is installed.
So true. And my point. And it doesn't matter who did it either. And
installs are solely modifications to storage. Just changes to some data
structures (the file system). So rather than argue those points (which it
seems we now agree on them) explain why we should describe these
modifications differently for different target systems? Why the information
about how these structures are interrelated is uninteresting past install?
Why versioning of applications (and their related structural changes)
doesn't have to be explict? And why should we do installs from within the
OS? Why not outside the OS where we can manage everything, across
platforms, in the same way?
> You are begining to sound like a person how has come up with a half baked
> idea to solve a problem that does not exist and is now trying to convince
> others that they have a problem so that your can sell them your solution
for
> it. Yes sir you have trouble right here in River City!
Well, we do. You might be managing a single system, but I am not. And I am
in the position of having to explain to customers how they are going to
manage their systems after I have gone to my next job. Or explain to them
that we have spent three weeks getting an application to work, but we don't
have anything to check in to the source code control system. ("We were not
working on code. It took us three weeks to get X to talk to Y, and for Z to
come up configured to point to A through firewall B. We are documenting it,
so if something goes wrong, you will be able <cough, look at toes> to
duplicate what we had to do to make it work. We will check that docuement
in. >")
------------------------------
From: "paul snow" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.text.xml,comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux, XML, and assalting Windows
Date: Mon, 28 Aug 2000 05:34:22 GMT
Some of these things have to be configured together. And I need them on
platforms other than NeXT and Mac. And I need to gather all the decision
points made during installs, and perhaps coordiante them with other installs
on different machines.
Sure, it isn't clear from the list I gave you. But I could have listed
Websphere advanced, with an Oracle adaptor, MQSeries client and server, DB2
server and client, along with DB2 Everywhere for our handhelds, Version X of
the C++ compiler, the proper JDK, DCE server, and DCE clients, etc. across
four NT Boxes, the firewalls (with the proper ports punched, and the ORB
configured to use them).
Simple installs (Such as NeXT-style bundles) would be a huge improvement
over what I generally have to deal with. Still, it wouldn't resolve the
bigger picture.
Today there is a big push to establish large abstractions that simply make
computing a homogenous place. I don't think that is really the answer.
Abstractions only do their work if they are properly deployed.
Craig Kelley <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> "paul snow" <[EMAIL PROTECTED]> writes:
>
> [snip about ditching "package managers" all together]
>
> > Ah those simple days of yore! But we can't and we won't go back.
> >
> > Today we have Java VMs and Adobe Acrobat viewers, and browsers, and
browser
> > plug-ins, word processing packages, and stock tickers, Internet based
games,
> > etc. Never mind that we are going to be configuring systems to connect
with
> > other systems, and use databases, and database clients, and we need to
set
> > up security, and down load the new versions of our clients, etc.
>
> ... enter NeXT-style bundles. They can handle all the above, and you
> only need to know how to use cp(1) to install them.
>
> > It isn't going to be simple in the future. It is going to get worse.
In
> > another post I list a set of requirements we are going to need from a
> > package manager. Not want, need. Typewriters are out for good. And
> > Redmond may be at fault to some degree, but if so they only pushed us
ahead
> > in time a bit. It was going to happen to us anyway.
> >
> > We have to have package managers, but they need to be based on open
> > standards. And they need to operate in an environment outside the
execution
> > environment of the supported computer systems. They need to be able to
> > manage cross platform, distributed applications. Why? Because we are
on
> > the Internet already! We want to bank, we want to order hamburgers on
the
> > Interstate Hwy so I don't wait for my order! I want to use my PDA to
adjust
> > my lights in my hotel (cause I don't know where the switches are, but I
have
> > my PDA), I want to listen to my MP3 files on the rental car's stereo,
from
> > the station I programmed on the Internet.
> >
> > We can't do all of this by coping all our files onto our bin directory.
> > Sorry.
>
> Why not? Isn't the operating system supposed to do the mundane
> things for us anyway?
> --
> The wheel is turning but the hamster is dead.
> Craig Kelley -- [EMAIL PROTECTED]
> http://www.isu.edu/~kellcrai finger [EMAIL PROTECTED] for PGP block
------------------------------
From: "David .." <[EMAIL PROTECTED]>
Subject: Re: how to block Ctrl+Alt+Del
Date: Mon, 28 Aug 2000 00:29:17 -0500
Vlad Nova wrote:
>
>
> yea - thanx - you're right:
> # What to do when CTRL-ALT-DEL is pressed.
> # ca:12345:ctrlaltdel:/sbin/shutdown -t1 -r now
>
> I could change it:
> ca:12345:ctrlaltdel:echo " ... off !!!"
> but how to activate changes (without reboot of course ;-) ? Does init
> def_level work ?
To have /etc/inittab re-read do:
/sbin/init q
--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter. http://counter.li.org
ID # 123538
------------------------------
From: "paul snow" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.text.xml,comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux, XML, and assalting Windows
Date: Mon, 28 Aug 2000 05:50:27 GMT
> You do realize that is software that gives any meaning to the data on the
> storage device beyond just a collection of bytes or bits arranged as fixed
> sized records in th case of DASD devices. It is the processor that gives
> any meaning to the contents of ROMs, without interpretation the data is
just
> a random collection of whatever method is used to record the data. Just
> like writing is just a random collection of marking for someone who can
not
> read it.
>
> What do you mean by "other platforms"? It seems that you are not using
the
> same meaning that is attached to it by the reset of us.
My point is that the storage defines the software. Meaning is another
issue, related closely to Magic.
The storage exists logically outside the execution environment of the
software. Storage can be constructed for a processor with the software we
want that processor to use.
platform -- a physically different computer or a different OS or a different
processor, or smart storage. I'm pretty flexible with how you understand
the term platform. Most readings work, best I can tell. I'll even take a
virtual machine or two.
BTW, Does it hurt to be reset ? ;-)
------------------------------
From: Geert Haustraete <[EMAIL PROTECTED]>
Subject: Re: Belgian keyboard layout in Suse 6.4
Date: Mon, 28 Aug 2000 06:00:08 GMT
In article <8o3v0m$dhr$[EMAIL PROTECTED]>,
> It seems that the belgian keyboard layout is not
> really belgian : I had the same problem.
> However I found the correct layout.
> In Yast, I chose azerty/azerty, which seems to
> work properly.
> To verify that you have the same layout as I have,
> you can logon as root, and type the following:
> # cd /etc/rc.d
> # ./kbd status
> Keymap
> /usr/lib/kbd/keymaps/i386/azerty/azerty.map.gz is
> loaded.
>
> This keyboard layout however is not used in X. To
> configure this, start SaX and select as keyboard
> layout:
> Generic 104-key PC
> Belgian
>
> The only character that is still wrong, is the �
> (trema, Umlaut) : it is not configured as a dead
> character.
>
Marc,
Thx, after many tries I also found out that the keyboardmap in X is not
the same as in the console. I wonder why that is but after
reconfiguration in XF86Setup it works fine now except the � as you say.
Geert
--
Thx a lot.
Geert Haustraete
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.text.xml,comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux, XML, and assalting Windows
Date: Sun, 27 Aug 2000 23:02:49 -0700
Reply-To: <[EMAIL PROTECTED]>
paul snow <[EMAIL PROTECTED]> wrote in message
news:2_lq5.20248$[EMAIL PROTECTED]...
> <[EMAIL PROTECTED]> wrote in message
> news:8o9s06$c3b$[EMAIL PROTECTED]...
> >
> > paul snow <[EMAIL PROTECTED]> wrote in message
> > news:%pTp5.18774$[EMAIL PROTECTED]...
> > >
> > > Oh, so all those hours I spent installing stuff on Solaris was really
> > > Windows?
> >
> > Does Solaris use Microsoft Windows terminology that you have used in
this
> > thread? If any of us have made a error as the the platforms you are
used
> to
> > using, it is as a result of the terminology that you have selected to
use
> to
> > present your ideas.
> >
> > > The point is that we need to get over the idea that installing is part
> of
> > > the abstractions that the OS provides. That mindset prevents us from
> > > developing technologies (such as those I am describing here) that can
> > > install across platforms.
> >
> > Technologies that already exist, assuming that the latest permutation of
> > your position is what you are really proposing.
> >
> > The are many things that don't map well or at all from one OS to
another.
> > How would you handle those details?
> >
> > Consider file and directory attributes and permission settings. They
> don't
> > map from one OS to another very well. The installation process needs to
> be
> > OS specific in this and in many other areas.
>
> Yes, we have developed a number of great tools for describing in abstract
> information, and rendering specifically to different targets. TeX, SGML,
> XML, and PostScript to name a few.
>
> Assume that I (as a developer) have two OS targets.
>
> One is a Forth based OS that still uses block addresses and has no file
> system at all. (The old Forth arguements against file systems: "File
> systems are bad, cause performance problems, and real programmers don't
need
> them. Just give us the blocks off the disk when we ask for them!")
>
> The other is Linux.
>
> Can we get any more different than that? So in my application's XML, it
has
> a <Forth> section that lays out each <Block number="1075"> and its
contents,
> and continues for every block it needs. My application also has a <Linux>
> section that details the directories, files, etc. that the Linux
> implementation requires.
>
> If it is so easy to see how I can describe these two installations, why is
> it so hard to believe we can cover Linux, Windows, Solaris, etc.?
>
> This thread isn't about auto-magically constructing valid represenations
of
> applications for any OS. This is about being able to describe a valid
> representation on any OS.
>
> Given a technology like XML that can be used to describe structured data
> (and variations of that structured data), use that to describe the set of
> software components for a computer system. Then render those components
as
> described, by managing in context their needs as defined as how they
should
> be expressed in storage.
>
>
What I see is that you fail to answer questions and keep changing your focus
when you stated goals are shown to be flawed. You have made many claims of
what should be but you have been more mercurious than hermes. You have used
empty phrases and invalid terminology to describe your goals, what ever they
are.
How about providing us with a concrete example of what you are proposing?
It does not have to be something big, just something simple like the "touch"
or the "yes" programs. Show how you would encode it to "render" the program
and install it on Linux (all platforms supported by Linux), BSD (pick you
favorite BSD), Solaris, MS-Dos, Windows NT, Windows 95, Forth, and to make
things interesting PET, CBM, TRS-80, Apple ][, Apple ][+, Series 1,
System 34, System 36, and an DOS/VS on an IBM 360 and OS on a IBM 370.
------------------------------
From: [EMAIL PROTECTED] (Darren Winsper)
Crossposted-To: alt.os.linux,comp.text.xml,comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux, XML, and assalting Windows
Date: 28 Aug 2000 06:14:04 GMT
On Sat, 26 Aug 2000 17:58:55 -0700, [EMAIL PROTECTED]
<[EMAIL PROTECTED]> wrote:
> > And they need to operate in an environment outside the execution
> > environment of the supported computer systems. They need to be able to
> > manage cross platform, distributed applications.
> Just what do you mean by that? It sound like a lot of empty talk devoid of
> real meaning.
I wonder if he realises you can practically do that with Mozilla now.
Take Chatzilla for example; you click on the "install Chatzilla"
button and Mozilla takes the .xpi file and installs it. It's platform
independant and easy. If Mozilla were to grow into a desktop shell,
then this ideal could be at least partially realised.
--
Darren Winsper (El Capitano)
ICQ #8899775 - AIM: Ikibawa - MSNIM: [EMAIL PROTECTED]
Stellar Legacy project member - http://stellarlegacy.sourceforge.net
DVD boycotts. Are you doing your bit?
This message was typed before a live studio audience.
------------------------------
From: Eric Carlson <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.install,linux.redhat
Subject: Re: changing X-window appearance
Date: Mon, 28 Aug 2000 02:53:48 -0400
As far as I know, you can't change the graphics extensively from X, you
can only fine-tune them. To redo a misconfigured X configuration, get
out of X (hit ctrl-alt-F1 and log in) and run Xconfigurator. This will
take you through the process you went through during install of choosing
the monitor and card, etc. When it gets to "start x when computer
boots" say NO! It's slightly more inconvenient to start x manually, but
if your X gets messed up, it's a pain to recover when you can't boot to
text. Anyway, when you get a config that looks good when Xconfigurator
tests it, press alt-F7 to go back to X, and press ctrl-alt-bksp to
restart x, hopefully it should look better than before. BTW, if you
take my advice about booting to text, when you start your computer
you'll log in as root, then type "init 5;exit". This switches to the
5th run level (graphics) and logs root out of that text login when the
switch is complete.
[EMAIL PROTECTED] wrote:
>
> Hello all,
>
> Recently I installed Win98 and Redhat6.0 on my PC. The Win98 looks as
> it should be, but the X-window looks a lot worse than it should - the
> resolution is quite low, the icons looked awkward and not the right
> size. I reckon the problem is that I did not configure the graphics
> card properly when installing Redhat6.0. So how shall I change the
> settings for the graphics card within X-window?
>
> thanks in advance
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
------------------------------
From: "Adam H." <[EMAIL PROTECTED]>
Subject: Recommendations on a good book.
Date: Mon, 28 Aug 2000 17:32:14 +1000
Hi,
I'm looking for a good linux book. (On either Redhat or Debian). I haven't
started to work on Debian yet - but I may move to it. (I've heard it's
better
than RH).
Anyway - I coudn't care less about the GUI. What I really need to know are
things
like:
* Setting up linux to route other computers to the internet via one modem.
* Setting up linux as a Dialup Server and provide IP addresses. (Like a ISP)
* Setting up POP mail access using something like Qmail (and the ability to
fetch it from another POP mail account).
* Setting up different html / ftp protocols and accounts
* Setting up firewalls.
Is their a good book, that explains all this thoroughly?
TIA
Adam
------------------------------
From: [EMAIL PROTECTED] (Daryl Fonseca-Holt)
Subject: Re: X windows
Reply-To: [EMAIL PROTECTED]
Date: Sun, 27 Aug 2000 09:21:00 -0500
On Fri, 25 Aug 2000 21:13:36 GMT, [EMAIL PROTECTED]
<[EMAIL PROTECTED]> wrote:
>I would like to run multiple X window sessions on my RH6.2 System..how do it
>do it....
>
I use this script called 'x' to start X windows. It figures out what is
available and starts the X session on it:
#!/bin/sh
if [ $# -eq 0 ] # check to see if arguements are given (color depth)
then a=24 # default color depth
else a=$1 # use given arguement
fi
if [ $a -ne 8 -a $a -ne 16 -a $a -ne 24 -a $a -ne 32 ]
then
echo "Invalid color depth. Use 8, 16, 24, or 32."
exit 1
fi
for display in 0 1 2 3 4 5 # checks for open display, starts X on next available
{
if [ ! -f "/tmp/.X$display-lock" ]
then
exec startx -- :$display -depth $a
exit0
fi
}
echo "No displays available."
exit 1
------------------------------
From: [EMAIL PROTECTED] (Michael Perry)
Subject: Re: debian potato - selecting 'profile' to install now gone?
Date: Sun, 27 Aug 2000 21:42:49 -0700
Reply-To: [EMAIL PROTECTED]
On 21 Aug 2000 21:02:29 GMT, Colin Watson <[EMAIL PROTECTED]> wrote:
>Leopold Maurer <[EMAIL PROTECTED]> wrote:
>>I've tried finding an answer to this problem with deja and on the
>>debian-site... couldn't find anything and now I ask for your help:
>>
>>I try to install debian 2.2 and everything runs fine, except that i'm
>>missing the option to install profiles instead of packages/dselect. I would
>>like to install a profile, as this seems easier than going through the
>>packages.
>
>I'm not sure exactly what way the installer looks now (it's some time
>since I did a Debian installation from scratch rather than an upgrade),
>but in potato the profile system has been partially replaced by various
>"tasks". Running 'tasksel' from a root shell after the reboot should (if
>tasksel actually gets installed, that is, which I'd hope it would) let
>you install those task packages.
>
>*rummage* See:
>
> http://www.debian.org/releases/stable/i386/
> ch-init-config.en.html#s-preselections
>
>(recombine onto one line, of course). This actually says that profiles
>are still available - perhaps I ought to try installing Debian on a
>scratch partition. :)
>
>--
>Colin Watson [[EMAIL PROTECTED]]
>"Racism is generally the last refuge of the unimportant."
My recent experience is on one system that Debian rebooted to an active
apt-get session. I have gotten out of the dselect stuff as early as
possible and then apt-got stuff from the basic system. This took me a
while; but it was fun :)
--
Michael Perry
[EMAIL PROTECTED]
==================
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.setup) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Setup Digest
******************************
