Linux-Advocacy Digest #913, Volume #28 Tue, 5 Sep 00 04:13:03 EDT
Contents:
Re: Sun cannot use Java for their servers!! ("Simon Cooke")
Re: Sun cannot use Java for their servers!! ("Geert Deweerdt")
Re: How low can they go...? ("Simon Cooke")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: ZDNet reviews W2K server; I think you'll be surprised.... ("Stuart Fox")
Re: How low can they go...?
Re: How low can they go...?
----------------------------------------------------------------------------
From: "Simon Cooke" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.lang.java.programmer,comp.os.ms-windows.nt.advocacy
Subject: Re: Sun cannot use Java for their servers!!
Date: Tue, 05 Sep 2000 07:16:51 GMT
"Zenin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> There are many levels of computer user. My parents, for instance,
> are very intelligent people, however their grasp of computers is
> rather limited. FrontPage isn't targeted as a solely "professional
> web publishing tool", it's also marketed as a simple tool for
> consumers to build their own "home page". It would not supprise me
> at all to see my parents use C:\ for their web site, after all "C:"
> is "their" hard drive, why shouldn't they be able to put whatever
> they like on it? And therein we find the real problem; My parents
> are nothing close to "stupid", but computer professionals they are
> not. As such, this is a bug, clear and simple, both in the OS (for
> allowing it in the first place) and the application (for exploiting
> it).
If you have such an issue with it, why are you letting your parents run a
Win9X operating system? Surely they'd be better of with NT 4.0 or Win2k,
where you can give them a user account instead of a root account and nicely
mollycoddle them so that you're the only person who can do anything with the
filesystem.
Simon
------------------------------
From: "Geert Deweerdt" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.lang.java.programmer,comp.os.ms-windows.nt.advocacy
Subject: Re: Sun cannot use Java for their servers!!
Date: Tue, 5 Sep 2000 09:17:51 +0200
Aaron R. Kulkis <[EMAIL PROTECTED]> schreef in berichtnieuws
[EMAIL PROTECTED]
> Stuart Fox wrote:
> > They have created a knowledge base article - Q183030 - available
publicly on
> > the web http://www.microsoft.com/technet. For customers who bought the
> > buggy version, they've provided a workaround (*don't* create Frontpage
webs
> > in the root of your drive - duh!), which is hardly an onerous
requirement.
> > It would appear the bug does not affect Frontpage 2000, as the article
only
> > references Frontpage 98 (although I'm not willing to test this :) )
>
> Translation: M% left the bug in place.
>
>
> --
> Aaron R. Kulkis
> Unix Systems Engineer
> ICQ # 3056642
>
Hi,
the discussion being a total newbie, I only wonna say this: I don't think it
was and is the intention of Microsoft to sell Frontpage to specialists. Au
contraire. So it is/was a bug.
Geert Deweerdt
------------------------------
From: "Simon Cooke" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Tue, 05 Sep 2000 07:29:20 GMT
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Simon Cooke in comp.os.linux.advocacy;
> [...]
> >So you don't understand the complexities of supporting 6 different OS's,
the
> >costs involved, and how much work that would be with today's
applications?
>
> No, I don't *care* about the complexities and work involved, because I
> don't have to do it. I'm more than prepared to pay for someone to do
> it, if necessary, *in a competitive market*. Competition is supposed to
> be really hard; its how we keep things efficient.
Okay... how about this scenario: What if the price of every piece of
software you bought increased by a factor of ~ 5, because it had to be
ported to 5 other OSes?
So... for example, Half Life would cost $250 a copy.
Would you buy it?
Probably not.
OK... so what if the company decided to support two or three really popular
OSs? ie. the most popular. They'd get the best return on their effort, and
sell more copies.
End result: eventually only 1, maybe 2 of those OSs still exist.
Now do you see why this argument is futile?
Is it any wonder that there used to be 20, maybe 30 different computers, and
now you've got Apple and Intel?
Have you observed what happens in the highly competitive console market? You
get three pieces of hardware, one of which dies in a year, one of which dies
in two years (or hangs on by a thread), and the other ends up with 90%
market share for four years. Rinse, repeat. I stopped buying consoles
because I didn't want to back a lemon and have to buy a new one.
End result though: for a while, one manufacturer comes out on top. Then the
cycle starts up again.
Artificially enforcing that there be 6 different main systems out there is
guaranteed to destroy the computer market. Heck, consoles can't even manage
it -- and they've tried.
The closest you could get to this situation is what happened with MSX. But
guess what? That's what a PC is today.
Simon
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:39:14 +0100
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Stuart Fox in alt.destroy.microsoft;
> >In article <8p0fmt$2gq$[EMAIL PROTECTED]>,
> > sfcybear <[EMAIL PROTECTED]> wrote:
> [...]
> >Once again you misunderstand Matt. What I suggested there was to use
> >the Unix version of BIND **only**, no Windows 2000 DNS at all. If you
> >run a reasonably recent version of BIND (8.1 from memory) you can do
> >this.
>
> Why would I have to run a recent version of BIND?
>
It needs to support
1. Dynamic Updates
2. SRV records
3. Incremental transfers (not sure if this is a requirement or just a
recommendation)
which older versions of BIND don't support.
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:42:22 +0100
"sfcybear" <[EMAIL PROTECTED]> wrote in message
news:8p1ctu$fp$[EMAIL PROTECTED]...
> In article <8p0n0o$a6c$[EMAIL PROTECTED]>,
> Stuart Fox <[EMAIL PROTECTED]> wrote:
> > In article <8p0fb3$26d$[EMAIL PROTECTED]>,
> > sfcybear <[EMAIL PROTECTED]> wrote:
> > >
> > > > >
> > > > > In many companies that would require a complete overhaul of DNS
> > and
> > > > > re-addressing of their workstations. Many places I have worked
> did
> > > > _not_
> > > > > all their MS stuff to one network, Unix and Mac to another. Poor
> > > > design,
> > > > > Maybe, but this is the REAL world and not everything is clean or
> > > well
> > > > > designed. The SysAdmin that forgets that is in for big problems
> > > > rolling
> > > > > out W2K's DNS.
> > > >
> > > > We do it with a couple of lines in the login script
> > > > if "%OS%=="Windows_NT" cscript ntdns.vbs
> > > > else cscript win9xdns.vbs
> > >
> > > ??? you miss understood... What if you have BOTH Unix and Windows
> > > _already_ on the same set of IP addresses with the SAME domain
> > > configured?
> >
> > What if? Big deal, you script the changes.
>
> That is IF you have good enough documentatation to DO that kind of
> scripting! many paces DO NOT!
What documentation do you require? How to write a script?
The documentation you need would probably be:
Which registry key to change (easy to find)
It's a standard location on each machine, it really is trivial.
>
> >
> > >A LOT of time needs to be spent RECONFIUGRING MANY installed
> > > clients! DHCP might help IF and ONLY IF you were already setup with
> > > DHCP! You keep talking as if you do have a large installed base that
> > is
> > > perfectly setup. that is NOT the case in many sites! Many sites have
> > > WIndows AND Unix on the same IP segment, on the same Domain with ALL
> > > info HARD CODED (no DHCP). It would be a nightmare to reconfigure
> the
> > > network the way you discribe! Why should I have to worry about what
> > OS I
> > > am running to set up DNS??? Shouldn't DNS be STANDARD????
> >
> > Where did I mention DHCP? I mentioned changing the DNS settings for
> > all Windows based hosts globally & unattended by a few lines in a
> login
>
> >
> > Like I say in another post, if you want to simplify the DNS, use BIND
> > 8.1 (?) or later on your *nix box to provide the DNS infrastructure
>
>
>
> Then why would we need to set up a second DNS server for W2K (which is
> the point I opbjected to).
To keep Unix admins off your back - who either a) don't want to give up
control of their DNS, or b) won't run a recent version of BIND.
>
> Form your post:
>
> "Pretty simple to design a system such that the Unix DNS can remain the
> root of the companies DNS - e.g. for the
> Win2K domain use
> nt.mydomain.com. That's exactly how we've implemented it, set the
> Win2K DNS up to forward to the Unix DNS, let
> the Unix DNS do what they like (forward, act as secondary - if they are
> running the right version of BIND), no
> problem."
>
>
> to different servers run by 2 different groups when that was NEVER
> required before???? Sounds like extra overhead to me! Not to mention all
> the work needed to change over (even your little scripts would be a pain
> if documentation is screwed up [as many paces are]
>
Documentation is not required - standard changes, standard addresses etc.
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:47:37 +0100
"sfcybear" <[EMAIL PROTECTED]> wrote in message
news:8p1dkn$12i$[EMAIL PROTECTED]...
> In article <8p0m84$993$[EMAIL PROTECTED]>,
> You did NOT claim that bind could do it all when you stated:
>
>
> "set the Win2K DNS up to forward to the Unix DNS"
Because that wasn't the issue I was addressing - unlike you, I stick to
topic. The issue I was addressing was how to provide a Windows based DNS
while still allowing Unix admins to control the root level domain for your
company.
>
> Why would you suggest the above then claim:
>
> What I suggested there was to use the Unix version of BIND **only**, no
> Windows 2000 DNS at all. Why would you want to set up a second set of
> servers when the bind emplementation will do just fine?
>
>
> why would you set up W2K to forward to Unix DNS if you telling me that
> you suggeted NOT using w2K DNS??? Please try to remember what you have
> said...
Like I also said if you bothered to read, was that there were many ways to
implement it, and what is easier for your company will probably be
different.
>
>
> And WILL BIND ENABLE ALL features of Active Directory???
>
Yes. DNS simply provides a locator service for the Active Directory. SRV
records are used to locate the various Active Directory service providers
(such as Domain Controllers), obviously A records (dynamically created by
Win2K hosts) are used to locate machines.
So as long as BIND supports
1. SRV records
2. Dynamic updates
3. Incremental transfers
it will enable all the features
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:48:55 +0100
"sfcybear" <[EMAIL PROTECTED]> wrote in message
news:8p1d4v$i2$[EMAIL PROTECTED]...
> In article <8p0mdh$9hf$[EMAIL PROTECTED]>,
>
> Yes I DO know how to do scripts and if you think it would work in a
> major corpration that has almost NO documentation, without major
> headachs, then go right a head and do it. And I the Unix man will sit
> and laugh as another ms admin bites the dust!
>
a) Most major corporations have documentation
b) You don't need documentation to write a script to change DNS entries
(unless you've forgotten where your DNS servers are) on Windows machines.
The change is trivial, so having documentation or not is irrelevant
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:49:55 +0100
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said Stuart Fox in alt.destroy.microsoft;
> >In article <8p0fst$2qb$[EMAIL PROTECTED]>,
> > sfcybear <[EMAIL PROTECTED]> wrote:
> >> > We do it with a couple of lines in the login script
> >> > if "%OS%=="Windows_NT" cscript ntdns.vbs
> >> > else cscript win9xdns.vbs
> >> >
> >> > Readdressing of workstations is a *minor* issue,
> >>
> >> Yeah if you only have 2 workstations! but what if you have
> >THOUSANDS???
> >> it becomes a MAJOR issue! This sort of BS is why I would NOT want to
> >> turn DNS over to a MS admin!
> >>
> >
> >Well a good Unix admin knows about things called
> ><drevil>"scripts"</drevil>. [...]
>
> A competent admin of anything knows that hand-waving something by saying
> "oh, just use a script" is the height of idiocy.
>
To use a script to update the DNS entries is trivial. It's not a
hand-waving explaination at all, it's a perfectly valid and simple solution
to the problem.
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: ZDNet reviews W2K server; I think you'll be surprised....
Date: Tue, 5 Sep 2000 08:56:33 +0100
"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Just to save you the effort, 'sfcybear'.
>
> Said Stuart Fox in alt.destroy.microsoft;
> >In article <8p0fb3$26d$[EMAIL PROTECTED]>,
> > sfcybear <[EMAIL PROTECTED]> wrote:
> [...]
> >What if? Big deal, you script the changes.[...]
> >Where did I mention DHCP? [...]
> >I can't see the nightmare scenario you suggest. [...]
> >As for DNS being standard, it really doesn't matter what you point at
[...]
>
> It seems so painfully obvious that Stuart is reading press releases
> instead of documentation that there's little reason to bother with him.
> And then he starts the backpedaling. The consistency of this technique
> is so god-damn baffling that I swear its a virus. Microsoft is a petri
> dish, and now its gotten out; even the amateur astroturfers know how to
> spin the 'ymmv' yarn to obscene proportions:
Of course, you've caught me out. I'm just a mere paper MCSE, who has never
seen a server before, and don't even know how to install Windows 2000. I
must congratulate you on your clever detective work. Or not.
The fact that I'm using this in production obviously hasn't occurred to you.
It's not based on press releases, it's based on what I've seen. It's Matt
that's changing his argument continously, I'm just providing different
answers to a changing argument.
To implement DNS in Win2K you have a few options, depending on what fits
your environment best. If you are running a recent version of BIND, you can
use the Unix based DNS to provide the infrastructure for your active
directory. If you're not, you can implement a Win2K based DNS. Or you can
use a combination, it really isn't that big a deal.
>
> >Like I say in another post, if you want to simplify the DNS, use BIND
> >8.1 (?) or later on your *nix box to provide the DNS infrastructure for
> >Windows & *nix. There's more than one way to skin a cat, whichever
> >fits your infrastructure best is going to be different for each
> >instance. Where I work, we simply set up a subdomain nt.mydomain.com,
> >forward everything else to the *nix DNS & no problem. I have yet to
> >see a problem with this.
> >
> >I still don't see why you're making this out to be such a big issue.
> >It would seem to me to be a lack of understanding on your part of the
> >different ways of using Windows 2000 DNS.
>
> That's a standard close, too. He used the same basic premise in the
> last exchange, where he said:
>
> >[...]The sort of BS
> >you're putting up is why most people have problems with WinNT - they
> >don't understand how it works.
>
> The problem, of course, with WinNT and W2K, DNS, et. al, is that it
> doesn't work. We seem to have a clue why, Stuart, and you don't. Would
> you quite prattling on about your wondrous technology where you 'just'
> rely on it to work in ways its known to fail?
Well it works here, so I guess I must be dreaming.
>
> It isn't because we don't understand it, Mr. Fox. It might be
> proprietary, but it ain't magic. Its because we do understand it, and
> know that understanding why it doesn't work isn't enough to make it
> work. Quite weaving your fantasies where more amateurs might become
> stupid from your example.
>
Which bit of what I have suggested doesn't work? DNS? BIND is standard.
Scripting a change to the DNS entries in Windows machines? Trivial, look
into WSH and you might learn something. Interoperating Win2K and BIND DNS?
Done it here (with the one *nix server that is running a recent version of
BIND)
Point out the errors in my statements, and I'll be glad to help you out,
otherwise butt out and take your drivel elsewhere.
> --
> T. Max Devlin
> -- Such is my recollection of my reconstruction
> of events at the time, as I recall. Consider it.
> Research assistance gladly accepted. --
>
>
> -----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
> http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
> -----== Over 80,000 Newsgroups - 16 Different Servers! =-----
------------------------------
From: <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Tue, 5 Sep 2000 00:04:33 -0700
Reply-To: <[EMAIL PROTECTED]>
Christopher Smith <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> In article <8p0t63$8us$[EMAIL PROTECTED]>, [EMAIL PROTECTED]
wrote:
> >
> >Christopher Smith <[EMAIL PROTECTED]> wrote in message
> >news:8outs2$liu$[EMAIL PROTECTED]...
> >
> >> You're kidding, right ? You seriously consider the current method to
be
> >> *worse* than the ones above ? The only one that anyone might consider
as
> >> better is the "trade a manual page for an upgrade".
> >
> >No I am not kidding, I am speaking with the experience of someone who has
> >lived through the times that these other mothods were predominate and I
have
> >worked with each of them from both the consumer and producer side of the
> >equation.
> >
> >The current method is the worse of the possibilities because of its has a
> >failure mode that is its critical weakness. The requirement of readable
> >distribution media from a pervious version of the product. All media
> >becomes unreadable with time. If someone has been upgrading every year
or
> >so using the current method for 10 years, then his media to prove his
right
> >to upgrade is 10-years or more old. When it it fails to be radable at
all
> >and perhaps the company responsible for the software is nothing but a
> >memory. How can the person again install the most recent version of the
> >software from its still readable media?
> >
> >Of course the best possible upgrade would be fair pricing. Price the
> >product to provide a fair profit and yet be low enough to be acceptable
to
> >both old and new users without having to resort to upgrade gimicks.
> >
> >If a fair price for a product would be $50.00, don't price it at $200.00
and
> >so that providing a $100.00 upgrade deal would seem to be a bargin.
> >
> >
What happened, an freak newsreader accident, or did my posting leave you
speechless?
IOW where are your new comments?
------------------------------
From: <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Tue, 5 Sep 2000 00:58:21 -0700
Reply-To: <[EMAIL PROTECTED]>
T. Max Devlin <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Said <[EMAIL PROTECTED]> in comp.os.linux.advocacy;
> >T. Max Devlin <[EMAIL PROTECTED]> wrote in message
> [...]
> >In fact, I didn't comment on RMI.
>
> It was Zenin. My mistake.
>
>
> [...]
> >Yes, both of these rpc services are functional, but could be better
> >implemented without using RPC and its overhead and other problem. I this
> >RPC is a good tool for some jobs but not for such critical systems. But
> >then, at the time that Sun first introduced these protocols were on an
RPC
> >kick..
>
> Well, they're fundamentally sound, well engineered, content-free
> transfer protocols; all RPCs are. While the name certainly makes it
> sound like there are some pre-defined 'procedures' which are remotely
> called, RPC is really just a connectionless session/presentation-layer
> wrapper around an almost entirely arbitrary application layer payload.
> How to take a message, and get it there, without any concern at all for
> what the message is. Very flexible, easy to re-use, but not what you'd
> call optimized or feature-rich.
>
> >If anyone wants to see how much of a performance hit that the RPC's
overhead
> >costs NFS. Setup a fileserver that offers a files set via NFS. Setup
> >marsnwe to provide that same files set on the same fileserver via NCP.
> >Transfer a few large files to and from the fileserver using both
protocols
> >and compare the transfer speed.
>
> How would that differentiate RPC's overhead from NFS'? This 'marsnwe'
> server doesn't use NCP on RPCs, I'm quite sure; that would be all but
> impossible. Or should I say disastrous.
>
> I thought lots of people had already essentially compared NCP to NFS (on
> RPC), and found that NCP was definitively faster. It wouldn't make much
> sense to just 'move' NFS to a different (new) transfer method, anyway.
> We'd want to redesign it. Make it more like NCP (but without the IPX
> shackles). ;-)
It would have to be a reimplemention from scratch using TCP/IP without using
RPC. The problem with RPC is that it tries to make the network calls
"transparent". To do this, the data passing through RPC has to be encoded
to be portable across the network. This adds processing overhead the encode
the data and then to decode it on the remote system. And the same is true
of the returned data but in reverse. The encoding of the data also expands
it size making it nessary to use larger network packets than is otherwise
needed or to transmit less real data in a single packet therby requiring
more network packets to transfer the same data.
More performance hit is taken by the fact that the program has to call the
rpc library to do it thing and the RPC library calls the socket library to
do it thing. If RPC were eliminated from the equation the program would
handle the socket library directly resulting in less code, smaller network
packets or packets containing more real data, all of which would mean better
performance.
I was not really concerned about the NCP except as a yardstick of how NFS
could perform if it did not use RPC. More performance gain could also be
possible by using TCP instead of UDP, but then perhaps some automatic
non-network fault recovery might be lost.
> So tell me more about this marsnwe.
marsnwe is the NetWare emulator than turn a unix box (both Linux or FreeBSD
are directly supported) into a NetWare fileserver as far as the client
machines are concerned. The emulations is not perfect in that not all of
the more exotic services are emulated and it is limited appearing as a
NetWare 2.x or 3.x box. It is better than real NetWare for many people
because it does does not restrict you disk volumes by the size of the
fileserver's RAM as real NetWare does. Then there is the cost, marsnwe is
free compared to paying thousands of dollars for NetWare, and marsnwe
supports some feature such as CD-Rom sharing that are extra cost addon
packages for NetWare. Performance wise, on an otherwise idle box is is a
little slower than NetWare 2.x but it is faster than NetWare 3.x, and MUCH
faster than NetWare 4.x on the same hardware.
A fileserver can offer the same fileset to various client using their native
filesharing protocols. You can run marsnwe on a filserver offering the
fileset via NetWare Core Protocol running on IPX/SPX. You can run Samba on
the same file server offering that same files using Microsoft Networking
running on TCP/IP. You can also offer the same fileset under NFS on the
same fileserver. The benefit of doing this is that a Dos box connecting to
this fileserver will find a NetWare fileserver. A Windows box connecting to
this fileserver will find a Windows peer offering shares. A unix box
connecting to this fileserver will find a unix NFS fileserver.
As for network printing from the non-unix clients, it is handled so that the
clients see that their native fileservers are handling their printing using
their native protocols. Behind the scenes the server software hands the
printing off to the unix printing system. Everybody sees what they expect
and everybody is happy. The are other protocols and package also available
to fileserver support for other platforms like Macs.
Fileservers that are being replace with these packages running on a unix
box, the fileserver is also availabe to do other work at the same time,
providing other network services, running batch jobs, etc.
Stability, this configuration has proven it stablity. On the order of
running one particuler fileserver as I have described here for over 4 years
and 10 months. It would have been longer if it were not for a black out.
------------------------------
** 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.advocacy) 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-Advocacy Digest
******************************
