Linux-Misc Digest #415, Volume #18               Wed, 30 Dec 98 23:13:13 EST

Contents:
  Re: Bash 2.02.1 (Affinity RPM) ("xming")
  Re: hylafax and print to fax (Bill Unruh)
  Re: When will kernel 2.2 be released? ([EMAIL PROTECTED])
  How to use Tape Backup ? (Doug Bryant)
  Re: Database Recommendation (Christopher Browne)
  Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance (Bill Unruh)
  Database Recommendation (Matthew Fleming)
  RedHat-5.2: ix86-libc5? ix (ishwar rattan)
  Re: Help, resolving IP address with a ppp conection (Bill Unruh)
  Primary Slave HD & FDISK (Ki-Won Lee)
  Re: Can redirect internal speaker beeps to soundcard? (AWing10651)
  Re: help me choose license (steve mcadams)
  Re: Any good charting tool? ([EMAIL PROTECTED])
  Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance (Mark Bashaw)
  CDRom installation problems (Prellwitz's)
  Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance (William Burrow)
  Re: Infringement of the GPL (Floyd Davidson)
  Re: Anti-Linux FUD (Victor Danilchenko)

----------------------------------------------------------------------------

From: "xming" <[EMAIL PROTECTED]>
Subject: Re: Bash 2.02.1 (Affinity RPM)
Date: 30 Dec 1998 21:54:01 GMT

I thought there are some incompatiblilities
betwee bash 1.x and 2x, try to searche the
gnu news archives

--
_____
xming
_____
REMOVE DOT COM to reply


Michael J. Saletnik wrote in message ...
>I pulled down the bash 2.02.1 rpm from Affinity Systems, advertised as
>a suitable drop-in replacement for RedHat 5.1. I definitely have
>wanted bash 2, so I installed it.
>
>I found that it broke /etc/rc.d/init.d/autofs such that I couldn't
>figure out how it broke it, just that it would complain on the ;; line
>of a case statement. My suspicions are that it has something to do
>with parsing )'s for subshell statements versus case statements.
>
>Anyway, has anyone else seen this happen? Was it violently stupid to
>replace /bin/bash with the newer version?
>
>-Michael
>--
>Michael J. Saletnik, PE  Tufts E'91 G'93 <[EMAIL PROTECTED]>
>Software Engineer, Thomson Financial Services
>Registered Professional Structural Engineer
>AIM: msaletni, ICQ: 24238794, www.tiac.net/users/icarus



------------------------------

From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: hylafax and print to fax
Date: 31 Dec 1998 02:24:18 GMT

In <[EMAIL PROTECTED]> Patrick O'Neil 
<[EMAIL PROTECTED]> writes:

>Stephen Richard FREELAND wrote:
>> 
>> : How does one print to fax from apps?

>I hope this isn't the case.  If I am working in Star Office on a doc
>that
>I need to fax to someone...they have to be able to see the text and in
>many cases the format must follow what I create on my screen.


print the document to a file using postscript. Then fax that postscript
file to the person using sendfax. hylafax looks at the file, sees that
it isi postscript, uses ghostscript to read the postscript file and put
out a tiff file to fax.
It looks "just" as though you had printed out the document onto the
printer and then sent that output as a fax.


------------------------------

From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.setup,comp.os.linux.advocacy,comp.os.linux.hardware
Subject: Re: When will kernel 2.2 be released?
Date: Thu, 31 Dec 1998 01:48:49 GMT

In article <[EMAIL PROTECTED]>,
  [EMAIL PROTECTED] (erikc) wrote:
> On 27 Dec 1998 17:43:24 GMT
> [EMAIL PROTECTED] (Filip M. Gieszczykiewicz) wrote:
> -- origin: comp.os.linux.hardware:
> >|In Article <[EMAIL PROTECTED]>, through puissant
locution, [EMAIL PROTECTED] (Harry McGregor) soliloquized:
> >|>Last I heard it the 2.1.xx kernels were up the M$ quality (about
> >|>125), and are not quite up to linux quality.  If you need things that
> >|
> >|Huh? My .96pl2 was up about M$ quality.... ;)
> >|
> >|(what scares the crap out of me is the Navy moving their fleet from
> >|[stable] unix systems to NT... [shivvvvvver])
>
> I read about that.  I just hope the Navy learned its lesson.  I
> believe the motivation for the switch over to NT was that the Navy
> wanted to get away from custom one-off hardware and software and move
> over to stuff they could get "anywhere".  Somehow, the brass seemed to
> overlook the the minor detail that the Navy's needs are one hell of a
> lot different than the needs of someone running a business.
>
> Maybe they should try Linux.  ;-)

http://www.ethepeople.com/etp/affiliates/national/fullview.cfm?ETPID=0&PETID=743
86&ETPDIR=affiliates/national

> has a petition you can sign if you want to nudge the gov't into
> considering open source software.

Actually that's a bit like trying to convices Linus Torvalds to use Linux.
Remember, many of the components which we refer to as the "Linux Distribution"
are/were funded through U.S. government projects including ARPA, Athena,
grants to BSD, FSF, MIT-Athena, NSF, and NCSA.  Much of the open source
software we have today was originally pulled from the SimTel-20 repository
which was Public Domain (any derivatives are legal).  In many cases, the
original authors transferred them from Simtel-20 to FSF/GPL.

It gets really funny when the government starts buying it's own software
from Microsoft, at $2000/head.

> Erikc (alt.atheist #002) | "An Fhirinne in aghaidh an tSaoil."
>                          |      "The Truth against the World."
>                          |                           -- Bardic Motto
> If we don't believe in freedom of expression for
> people we despise, we don't believe in it at all.
>    ---- Noam Chomsky
>


--
Rex Ballard - Open Source Advocate, Internet Architect, MIS Director
http://www.open4success.com

============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

------------------------------

From: Doug Bryant <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: How to use Tape Backup ?
Date: Wed, 30 Dec 1998 18:28:24 -0500

Hello all

I have become the proud administrator of a new linux server at work.
(the owner did not care what I used as long as it worked.  COOL )   I
got everyones accounts in the small workgroup up and running with samba
without a hitch.

I have been using Linux for a year or so but have never had to use or
had available a tape backup unit and I am not real sure how to use this
thing.   I played around with bru2000 but the non graphical version says
license expired. (The version included with 5.2)

I did not care to use the graphical version as I want to run the backups
at night as a cron job.

I found the program mt and can access the tape drive and rewind it,
erase it.   I do not really understand the purpose of moving around the
blocks (to the beginning of the first and end of the previous, etc.
Well I understand what it does but when and why does to do it)

I also found the dump program..  I played with this and backed up the
home directory as a test ( not very big yet )   the program kept asking
me if the next share was ready.

Anyway, is dump the best way to do this?  Any suggestions?  Better yet,
If anyone has any scripts that I could use that would be very nice.

Oh yea, the tape is a SCSI HP that uses 4Gb dat tapes.  I do not
remember the exact model.  (at home now)   The backup will only run on
the local machine, atleast for now.

Thanks for all the help.

Doug Bryant


------------------------------

From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: Database Recommendation
Date: 31 Dec 1998 02:55:12 GMT
Reply-To: [EMAIL PROTECTED]

On 30 Dec 1998 19:09:04 GMT, Matthew Fleming <[EMAIL PROTECTED]> wrote:
>I run my pathology lab with a customized database application running
>under Linux. The application was developed using the /rdb database
>package, which at the time I wrote it (a long, long time ago) was the
>only set of database tools available for Linux. The /rdb programs are
>accessed directly from the shell; the shell is essentially regarded by
>the /rdb developers as a kind of 4GL for this purpose. So the
>application consists mostly of a bunch of shell scripts.
>
>Anyway, I would like to replace this system with a new one based on an
>SQL database engine. If possible, I would like to access the SQL
>programs from the shell, since this would allow me to initially use
>many of my existing scripts.  Eventually I would like to build some
>Web clients using Java/JDBC, and to migrate many of the shell scripts
>to perl.

>The problem is that I am having a great deal of difficulty choosing a
>amongst the many database packages now available for Linux, and would
>like some recommendations.  

It's a tough call picking an SQL DB; no doubt you've seen the whole pile
of DBs at <http://www.hex.net/~cbbrowne/rdbms.html>.  I specifically
eschew providing overall recommendations.  You just have to look at the
jockeying that takes place surrounding TPC benchmarks to see that
claiming things about the performance of SQL DB systems is a dangerous
thing to do... 

I would ask the question of whether or not it is fundamentally necessary
to replace the application, which would certainly require, warrant, and
result in a substantial redesign effort, or whether it might be a good
idea to look at one of the "later" /rdb variants, and just spiff up the
existing application. 

If you're aware that there are 3 /rdb implementations out there, feel
free to ignore this; if not, you may want to look at one of the more
modern versions, as they might improve performance/functionality
without requiring a substantial application rewrite...

<ftp://ftp.rand.org/pub/RDB-hobbs/> is the "classic" edition that seems
to get updated now and again; 

<http://uu-gna.mit.edu:8001/uu-gna/schools/vslis/rdb> and
<http://www.branch.com/~rsw> should get you at the commercial /RDB
implementation, that claims extended functionality including X/GUI
stuff. 

You might be less likely to be aware of the most recent "NoSQL"
<ftp://ftp.linux.it/pub/database/RDB/nosql-0.8.tar.gz> implementation
that is written using awk, and claims to be faster than Rand /RDB due to
awk having a lower memory footprint than Perl.

<http://www.simtech-soft.com/dbops.html> is a system that was based on
/rdb that might be interesting from an academic standpoint if not from a
practical standpoint... 

-- 
"On the Internet, no one knows you're using Windows NT"
-- Ramiro Estrugo, [EMAIL PROTECTED]
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/lsf.html>

------------------------------

From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance
Date: 31 Dec 1998 03:04:46 GMT

In <[EMAIL PROTECTED]> Mark Bashaw <[EMAIL PROTECTED]> writes:

>Jeremy,
>You are correct in that most OS's store dates in the correct format
>"internally".  This is where we run into problems.  Unix, being a 32bit OS from
>long standing, uses a 32bit clock library that basically increments from an
>arbitrary date set in the early 1970's.  Unix doesn't actually know what day
>and time is, only that so much time has elapsed since January 1, 1973 (or
                                                                    ^^ 1970

>whatever), and the date is calculated accordingly.  This 32bit library will run
>into problems, I believe, in 2027 when it hits it's limit and rolls back over
                              ^^^^ 2038
>to 0.  At that point I  hope we'll be using a higher bit clock library or newer
    ^ -2 10^9 ( ie a date of 1901)


>OS rev's to alleviate the problem.

Except of course all those programmers who knew that time_t was a 4 byte
int and thus used int or long int in their programs rather than time_t.



>On the application side, in my Y2K research for my firm, we found that only
>programs which contravened correct programming practice may cause problems over
>the Y2K boundary, given that your hardware and OS are Y2K compliant.  An

Of course what "correct programming practice" is can change in time. In
1980 with 16K of memory ( or in an imbedded controller with 2 K memory)
using 2 byte years might be very "correct" 

------------------------------

From: [EMAIL PROTECTED] (Matthew Fleming)
Subject: Database Recommendation
Date: 30 Dec 1998 19:09:04 GMT


I run my pathology lab with a customized database application running
under Linux. The application was developed using the /rdb database
package, which at the time I wrote it (a long, long time ago) was the
only set of database tools available for Linux. The /rdb programs are
accessed directly from the shell; the shell is essentially regarded by
the /rdb developers as a kind of 4GL for this purpose. So the
application consists mostly of a bunch of shell scripts.

Anyway, I would like to replace this system with a new one based on an
SQL database engine. If possible, I would like to access the SQL
programs from the shell, since this would allow me to initially use
many of my existing scripts.  Eventually I would like to build some
Web clients using Java/JDBC, and to migrate many of the shell scripts
to perl.

The problem is that I am having a great deal of difficulty choosing a
amongst the many database packages now available for Linux, and would
like some recommendations.  I would like something that is stable,
well-documented, and has Java and perl drivers.  If the package had a
decent 4GL, I might use that as well.  The application is unlikely to
have more than a few concurrent users, so blinding speed and 
efficiency is not a requirement.  I would be willing to pay for a
commercial package, but of course the less money I have to spend, the
better. It is especially important that the package be well
documented, and that some form of support is available.

Thank you for your recommendations,

Matthew Fleming
==============================================================================
Matthew G. Fleming, MD                  phone : 414.456.4072  
Associate Professor                     fax   : 414.456.6518
Department of Dermatology               s-mail: Dept. of Dermatology
Medical College of Wisconsin                    Medical College of Wisconsin
                                                MFRC Room 4061
e-mail: [EMAIL PROTECTED]                             Milwaukee, WI 53226-4810
==============================================================================


------------------------------

Date: Wed, 30 Dec 1998 15:15:35 EST
From: ishwar rattan <[EMAIL PROTECTED]>
Subject:    RedHat-5.2: ix86-libc5? ix

Hello,
 
I just installed RedHat-5.2. Does it confirm to ix86-glibc or
ix86-libc5 library family?
 
Thanks in advance,
- ishwar rattan

------------------------------

From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: Help, resolving IP address with a ppp conection
Date: 31 Dec 1998 03:09:45 GMT

In <76d0ht$7s8$[EMAIL PROTECTED]> "Tom Kelly" <[EMAIL PROTECTED]> 
writes:

>However, I am unable to obtain a responces from any other name or IP
>address. Each time I try ping, I waits for about 3 mins and then reports

Sounds like a route problem. Type route and see where the default route
points to. It should point to the remote system at the other end of your
ppp connection. Run pppd with the defaultroute option (eg in
/etc/ppp/options)


------------------------------

From: [EMAIL PROTECTED] (Ki-Won Lee)
Subject: Primary Slave HD & FDISK
Date: 31 Dec 1998 03:05:18 GMT

Hello,

I've successfully installed a 2nd HD on my box as a Primary HD and my BIOS 
recognizes it fully.  It's a Quantum Fireball EX 6.4GB.  
This is the unit on which I'd like to dedicate Linux on.

The problem I'm having is that when I run Linux fdisk, it tells me that it
only has 784 cylinders instead of 1023.  I installed RH 5.0 anyway to see
what happens and the install went without a hitch, but when I do a "df -h"
on it, it tells me the followng:
/dev/hd1        100M    (this is my "/")
/dev/hd2        5.5G    (this is my "/usr")
/dev/hd3        61M     (this is my "/home")
and I have about a 100M of swap space accoring to my top.  (Please excuse
me if the output of df doesn't look exact, I'm going by memory here)

As you can see, it's only recognizing about 5.8G at most of my HD instead
of 6.4G.  The LBA is enabled for this HD as well in my BIOS and like I
said, it's fully recognized by it.  According to fdisk, I got this info:
255 heads, 63 sectors, 784 cylinders 
Units = cylinders of 16065 * 512 bytes.

Can someone please shed some light as to why fdisk is only recognizing
only part of my HD and not the fully 1023 cylinders?  During the install,
I chose to check my HD for errors and there were none.

With Best,

Steve

[EMAIL PROTECTED]
[EMAIL PROTECTED]

------------------------------

From: [EMAIL PROTECTED] (AWing10651)
Subject: Re: Can redirect internal speaker beeps to soundcard?
Date: 31 Dec 1998 02:23:48 GMT


I am a Linux newbie, and I haven't even gotten my sound card working, but why
would you want the beeps coming through your sound card?  I guess it might be
possible to cut the PC speaker wires and connect them to your sound card
somehow. . .


-Chris <[EMAIL PROTECTED]>

------------------------------

From: [EMAIL PROTECTED] (steve mcadams)
Crossposted-To: linux.redhat.misc
Subject: Re: help me choose license
Date: Wed, 30 Dec 1998 23:46:35 GMT

Thanks for your comments John; I've filed your posting away for later
reference.  See below for replies.

On Wed, 30 Dec 1998 20:36:51 GMT, [EMAIL PROTECTED] wrote:
>First you should understand that you as author are not bound by any license
>you publish under.  Releasing your library under the GPL does not prevent
>you from distributing the exact same code under any other license you
>choose.  Thus companies that want to use your library but do not want to
>put their code under the GPL can purchase proprietary licenses from you.

Thank you!  I was under the impression that once the code had gone out
under GPL that it was from then on "free for all" and that I, along
with everyone else, would not be able to include it in commercial apps
after it was GPL'ed.  This bit of information answers most (if not
all, I'll have to think about it) concerns I have had about GPL.  And
I would also be able to sell proprietary licenses?  Awesome.

Can you point me to sources where I can learn more about this aspect
of licensing and the GPL?

I guess that if someone else was involved in development, for example
on the Linux port, this would get more complex; I don't understand
just what the ramifications would be.

>  Please take the matter of licensing
>seriously, and please do not attempt to write your own open source license
>without consulting a copyright attorney who knows free software, or at
>least asking for help from knowledgeable people such as those on
>debia-legal.  Every week or so we see another homemade license that
>contradicts itself or fails to do what the author intended.  Crafting free
>software licenses is, unfortunately, damn near as difficult as crafting
>free software.  And not nearly as much fun.

I am already taking the matter seriously; the rest sounds like good
advice.  It's tough enough crafting a proprietary license, where all
you have to do is include lots of "you may not" phrases with a couple
"you may" phrases.

>I suggest that you use the GPL, with a notice to the effect that anyone who
>wants to see their patches accepted should grant you a non-exclusive
>unlimited license.  While it would be possible for someone to fork a
>version, there is no reason why anyone would use such a version as long as
>you are doing a good job.

Can you tell me more about this?  Any examples of existing licenses
like this which I can go look at?

>  If you die or go over to the dark side :) or
>move to Tibet and take up yak herding, the library will live on.

I doubt that I'll die (too soon anyway) or go over to the dark side,
but the yak herding scenario has a certain appeal...  Thanks.  -steve
========================================================
Tools for programmers: http://www.codetools.com/showcase

------------------------------

From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.x
Subject: Re: Any good charting tool?
Date: Thu, 31 Dec 1998 01:57:04 GMT

In article <[EMAIL PROTECTED]>,
  [EMAIL PROTECTED] (brian moore) wrote:
> On Mon, 21 Dec 1998 20:15:52 GMT,
>  [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
> > I need an easy-to-use charting tool (as opposed to drawing tool)
> > for Linux.   Visio is excellent but only runs on Windows.  Any
> > suggestions?
>
> Tried xfig?

Xfig is a general drawing tool.  I need something that can create UML
diagrams very easily... specifically collabaration diagrams.


============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

------------------------------

From: Mark Bashaw <[EMAIL PROTECTED]>
Subject: Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance
Date: Wed, 30 Dec 1998 15:55:49 -0800

Jeremy,
You are correct in that most OS's store dates in the correct format
"internally".  This is where we run into problems.  Unix, being a 32bit OS from
long standing, uses a 32bit clock library that basically increments from an
arbitrary date set in the early 1970's.  Unix doesn't actually know what day
and time is, only that so much time has elapsed since January 1, 1973 (or
whatever), and the date is calculated accordingly.  This 32bit library will run
into problems, I believe, in 2027 when it hits it's limit and rolls back over
to 0.  At that point I  hope we'll be using a higher bit clock library or newer
OS rev's to alleviate the problem.

DOS and it's derivitives, utilize a different starting date and smaller bit
library so the rollover problem occurs earlier.  Only OS patches and/or
upgrades alleviate the problem fully.

On the application side, in my Y2K research for my firm, we found that only
programs which contravened correct programming practice may cause problems over
the Y2K boundary, given that your hardware and OS are Y2K compliant.  An
example of incorrect programming practice would be to query the Real-Time Clock
or BIOS clock directly to get date information and use an internal clock
library, as opposed to querying the OS clock for date information.  No
applications that we checked did this kind of thing.  Most get their clock info
from the OS.  That's not to say that a user created script or database can't
have been written incorrectly, but that in most cases, the OS, and the
applications supplied with it, will function correctly.

Mark Bashaw


Jeremy Mathers wrote:

> Actually, DOS (and its derivatives) do, like Unix, store dates in Y2K
> formats (there may be exceptions, but this is true for the most part).
>
> The Y2K problem is mostly about old mainframe OS's and apps, but it also
> affects the external interfaces of modern OS's - such as the COMMAND.COM
> 'date' command.  Remember, the line between system and app is a grey
> one, and I wouldn't be at all surprised to find somewhere, in some
> "system" database or config script, a 2 digit date assumption in Linux
> or some other Unix system.
>
> And then, of course, there is all the user written stuff for Unix and DOS.


------------------------------

From: Prellwitz's <[EMAIL PROTECTED]>
Subject: CDRom installation problems
Date: Wed, 30 Dec 1998 20:00:44 -0600

    I'm fairly new to linux but I still know somewhat what i'm doing.  I
am trying to install Redhat 5.2 on an intel 133 with a 2 gig HD and 2
CDRom drives.  I have Win 95 installed but when I boot with the boot
disk and try installing, it says the CDRom drive is not mounted
Any help is appreciated
Ry


------------------------------

From: [EMAIL PROTECTED] (William Burrow)
Subject: Re: Linux (Red Hat 5.1 and 5.2) Y2K compliance
Date: 31 Dec 1998 03:37:26 GMT
Reply-To: [EMAIL PROTECTED]

On Wed, 30 Dec 1998 15:55:49 -0800,
Mark Bashaw <[EMAIL PROTECTED]> wrote:
>DOS and it's derivitives, utilize a different starting date and smaller bit
>library so the rollover problem occurs earlier.  Only OS patches and/or
>upgrades alleviate the problem fully.

DOS allows 7 bits for the year, with an epoch of 1980.  DOS is good until
2107, which is later than Unix's date rollover problem.  Long live DOS. ;)


-- 
William Burrow, VE9WIL  --  New Brunswick, Canada     o
Copyright 1998 William Burrow                     ~  /\
                                                ~  ()>()

------------------------------

From: [EMAIL PROTECTED] (Floyd Davidson)
Subject: Re: Infringement of the GPL
Date: 31 Dec 1998 02:56:25 GMT
Reply-To: [EMAIL PROTECTED]


Damien Kick  <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Matthew Malthouse) writes:
>
>> On 24 Dec 1998 21:17:59 GMT K. Spoon wrote:
>> } [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>> } >  Damien Kick writes:
>> } >  > I've been wondering lately if there have been any cases
>> } >  > brought before a judge regarding infringement of the GPL.
><snip>
>> } Not brought before a judge....  open source developers don't
>> } really have the money to afford the lawyers to go after people who
>> } steal code from the rest of the world.
><snip>
>
><nod> Someone mentioned getting an attorney to do the work pro bono
>but it is an unfortunate reality that money (or lack thereof) can win
>a lawsuit in and of itself.  How often do parties settle lawsuits out
>of court without regard for actual guilt or innocence just to avoid
>the cost of going to court?

Indeed.  It is well to remember that pro bono means the
attorneys do not charge for their time, but that does NOT mean
an attorney pays for the expenses!  The various costs, from
copying, to filing fees, to travel expenses and the hiring of
expert services for investigation, research, analysis, or what
ever, are often at least as significant as the actual cost of
legal services.

Law suits are expensive.

  Have a good day,
  Floyd

-- 
Floyd L. Davidson                                [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska)                       [EMAIL PROTECTED]
  Pictures of the North Slope at  <http://www.ptialaska.net/~floyd>

------------------------------

Date: Wed, 30 Dec 1998 16:46:16 -0500
From: Victor Danilchenko <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.x,uk.comp.os.linux
Subject: Re: Anti-Linux FUD

Evan Carew wrote:
> 
> Victor Danilchenko wrote:
> 
> It might be a small point to you, however, It has been a huge point to the networked
> community for quite some time. The source of the problem is that it is always 
>cheaper to
> share networked resources than to duplicate them on each user's node. In any large 
>network
> I'm aware of, any time the server goes down, the user is down, even if we are only 
>talking
> M$ products.

        Anyway, the network resources are certainly shared; however, please
give me one good reason to remove a local copy of tcsh in favor of
network copy which reads a network init file.
        Nobody is talking about replicating work -- we do not install a
separate teX and gcc installations on each system. However, the thread
is about shell initialization files...

        So, how about have a 10-day experiment continue to run uninterrupted?..
        Yes, they would certainly be unable to use any network features if the
network was down; however, they will still be able to continue doing
much of the stuff they were doing before. Previously build expreriments
will still execute, the shell will still work, editors will still run,
etc.
        I don't know how things stand in your network, but in our network, if
the server is down, ONLY the things that were on the server are down.
Each system can still be used as a standalone node -- it's not very
useful, but you can still write TeX code on a standalone system, for
example, or generate the output of your reinforcement learning
experiment, and being able to spend network downtime doing that is
better that spending the same time twiddling your thumbs. However, if
the system gets its shells from NFS, you won't even be able to do that
much...

-- 
|  Victor A. Danilchenko       CSCF support  |
|  [EMAIL PROTECTED]       A313, 5-4231  |
+--------------------------------------------+
|       Quando omni flunkus, moritati.       |

------------------------------


** 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.misc) 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-Misc Digest
******************************

Reply via email to