Linux-Advocacy Digest #129, Volume #28 Mon, 31 Jul 00 12:13:05 EDT
Contents:
Re: Micro$oft retests TPC benchmark (fungus)
Re: C# is a copy of java (Donal K. Fellows)
Re: Micro$oft retests TPC benchmark (Donal K. Fellows)
Re: Would a M$ Voluntary Split Save It? (Leslie Mikesell)
Re: Micro$oft retests TPC benchmark ([EMAIL PROTECTED])
Re: Aaron Kulkis -- USELESS Idiot -- And His "Enemies" -was- Another (Roberto
Alsina)
Re: The Failure of the USS Yorktown (Donal K. Fellows)
Re: Anti-Human Libertarians Oppose Microsoft Antitrust Action (Roberto Alsina)
Re: Micro$oft retests TPC benchmark ("Mike Byrns")
----------------------------------------------------------------------------
From: fungus <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Micro$oft retests TPC benchmark
Date: Mon, 31 Jul 2000 15:09:40 GMT
Mike Byrns wrote:
>
> Everything is redundant.
>
No it *isn't*, that's the entire point. How many times must
it be pointed out to you?
The Microsoft TPC database is split into equal chunks, if
any chunk/machine is lost then all processing stops for that
part of the database.
If part of a big (eg.) Sun server fails all parts of the
database keep on running.
If you really can't see the difference then you're beyond
hope.
--
<\___/>
/ O O \
\_____/ FTB.
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Subject: Re: C# is a copy of java
Date: 31 Jul 2000 15:14:31 GMT
In article <8m08f1$h9k$[EMAIL PROTECTED]>,
<[EMAIL PROTECTED]> wrote:
> More runtime overhead slowing the performance for the end user. Besides
> there are many cases where a garbage collector can not prevent memory
> leakage any way. Not to mention that evey dereference of dynamic memory
> would be twice as slow and and assignment to pointers would not be performed
> in O(n) time. Your programs would consume more memory for garbage
> collector house keeping as well.
It depends on how the GC is implemented; there have been some *very*
bright people working on it for many years. (Assignment to pointers
is an O(1) operation!)
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- I may seem more arrogant, but I think that's just because you didn't
realize how arrogant I was before. :^)
-- Jeffrey Hobbs <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Micro$oft retests TPC benchmark
Date: 31 Jul 2000 15:01:32 GMT
In article <Uagh5.9644$[EMAIL PROTECTED]>,
Mike Byrns <[EMAIL PROTECTED]> wrote:
> The "minimize points of failure" rule only applies to
> infrastructures that are non-redundant.
*Thwap!* No. Clusters are used *because* they reduce the number of
points of failure. However, if your software immediately undoes this
by putting different critical data on each system in the cluster, you
have managed to go from no single point of failure (except maybe the
network itself) to *every* system being a single point of failure.
Which plays merry hell with your reliability and availability.
> This is a cluster we've been talking about. Everything is
> redundant.
With statements like that one above, the main redundancy round here to
be hoped for is yours. Stop that slack-witted drooling! It is not
edifying in a supposed computer professional...
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- I may seem more arrogant, but I think that's just because you didn't
realize how arrogant I was before. :^)
-- Jeffrey Hobbs <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Crossposted-To:
comp.os.os2.advocacy,comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: 31 Jul 2000 10:28:36 -0500
In article <NkFg5.13447$[EMAIL PROTECTED]>,
Daniel Johnson <[EMAIL PROTECTED]> wrote:
>> Which tool is it that Novell needs to be able to write a
>> replacement for Active Directory that will fully interoperate
>> with everything else MS is shipping?
>
>One option is Microsoft Visual C++.
And a large supply of unavailable trade secret information.
>Admittedly, that is only going to work if you want your
>new Active Directory clone to run on Windows of
>some sort. For NetWare, I don't know- but I assume
>Novell can get compilers for their system.
>
>You also need MSDN, but that's available on the web.
And it will lead you astray by suggesting that you
can't actually interoperate but instead have to
replace the client to match the server.
>No need for the protocols; MSDN tells you how to do
>it.
Where does it tell you how to interoperate fully with
the existing unmodified clients?
>[snip]
>> >Then again, you may. Certainly MS isn't standing in the
>> >way.
>>
>> I'll believe that when I see fully compatible Active Directory
>> replacements.
>
>May I ask in particular what you feel the shortcoming of
>the current Novell NetWare product is, in this role?
I am only going by someone else's report that they
cannot duplicate the functionality completely. I assume
it is in the proprietary addition to the kerberos mechanism.
>[snip]
>
>But what protocol would you use to talk to the database
>on that IBM 390? WIth Windows I would install the ODBC
>driver for it, and would be blissfully ignorant of the protocol
>it used.
You can only remain ignorant of a protocol when it works. If
IBM provides the DB2 ODBC driver and it works correctly
across the platforms that is find. But when MS provides
something claiming to be PPP, or HTML, or java byte-code
and it doesn't work correctly across platforms, then
you can't remain blissfully ignorant.
[snip]
>> Each one provides only it's own implementation of the
>> cross-platform standard. One platform, one program,
>> regardless of the number of other endpoints.
>
>That's nice, but it demands you be able to change each
>platform to the chosen standard. With plug-ins, you can do
>this, but on some platforms it won't happen.
That's the point of standards. Non-standard proprietary
protocals exist to prevent interoperation, so of course
you won't be able to use them everywhere.
>[snip]
>> If you do internet networking right, you automatically
>> do local networking right. Why do it twice?
>
>Doesn't work. LANs aren't just tiny bits of Internet.
>They are used for different things.
It doesn't matter at the ethernet packet layer and it
doesn't matter at the next layer up either. IP is
a fine transport for lan packets.
>Being able to do web pages, as on the Internet, doesn't
>give you database access, for instance.
The same LAN configuration can give you both - if you
use IP in the first place. Or it can require you
to do everything twice if you don't.
>This is
>why there are such things as "virtual private networks",
>which layer over the Internet and make it do things like
>LANs do (except, of course, not so local anymore)
This is also much harder to do if you insist on multiple
protocols in the encapsulated layer. With IP only it becomes
a simple matter of address translations, and optional
encryption.
>[snip]
>
>> For the 'got them'
>> to happen, the customers have to be deceived.
>
>Not at all. They have to be convinced to use the extensions,
>but the easiest way to do this is to make those extensions
>useful.
No that would be, ummm... useful, and then it wouldn't
be necessary to tie the extensions to an existing protocol
and pretend they are following standards.
>Remember, most people do not consider "standards compliance"
>as a goal they want to acheive. It's just not the point.
Who doesn't? It shouldn't have to be the point - something
claiming to follow a standard should, so there should be
no question about it.
>
>> >Microsoft has the insight to see that it doesn't matter
>> >if you *also* support the 'open' protocols. You've
>> >still got them.
>>
>> Only if the open protocols aren't quite usable,
>
>No, even if they are usable it still works.
Only if the product is actually better than any of the
alternatives that can easily be substituted. That is
rarely the case, and it is really the need for proprietary
extensions that were never documented as being non-standard
in the first place that keep you locked in.
>> and they
>> have carefully made sure of that and made it difficult
>> to tell what is standard and what is a MS-specific
>> extension.
>
>While MS doesn't go out of their way to point out the
>differences, it's certainly possible to discover them
>if you care.
>
>Most don't.
Everyone cares, but you are right that most don't discover
the truth until it is too late.
>That's why this works.
Yes, it is extremely deceptive.
>> XML/XSL is the hot new thing. The standards
>> have nothing to do with Microsoft, and certainly nothing
>> to do with COM. But, poke around the examples that MS
>> provides and I suspect you will find that they show COM
>> objects as the way to do things - follow them and you might
>> as well forget that there is a cross-platform standard.
>
>Darn straight. MS, after all, wants to document *their* tools,
>not somebody else's.
And again they want to claim standards compliance while
deceiving users into non-compliance.
>> And they still don't do a standards-compliant XSLT.
>
>What XSLT?
The one in IIS (server-side) and IE5 (browser side). They
released before the standard was finalized so I'm willing
to believe that their motives were not necessarily evil
to start with. However the standard was set in November
and we still don't see a compliant version from MS, so they
obviously have no interest in conforming at this point.
>[snip]
>> If the nature of the protocol allows a transparent transformation
>> between the open standard and a (perhaps more efficient) proprietary
>> version then it doesn't hurt to use the proprietary version
>> where it works.
>
>Ah, there's the trick to it: MS's "extended" standards do work
>like this: they degrade to the "open" version when necessary.
If that were the case, j++ generated byte-code would run
in a standard JVM, and all Frontpage-generated HTML would
display in a standards-conforming browser. But that isn't
the case. The MS extensions are designed specifically
to not allow interoperation with competing products.
>You were yourself complaining about the effect: MS customers
>see that it "doesn't work as well" when they connect to Unix
>(or whatever) and want Windows Everywhere to correct. They
>don't realize the extra features they are used to aren't part
>of the Standard, and wouldn't care if they did know.
That isn't the case either, because fully standards-conforming
versions of things work just fine. What they don't realize,
at least at first, is that they have been deliberately locked
into using a single vendors products because even though it
claimed standards conformance, it didn't deliver.
>> >> Is it implemented for your target platform?
>> >
>> >This problem is no less great with 'open standards'.
>>
>> Examples?
>
>Ever tried to talk to a System/36?
>
>'Tis loads of fun.
Do they still make those? If they do, I'd expect IBM to
supply standards-compliant software or at least a programming
environment capable of building your own. If not, the
issue isn't relevant.
>[snip]
>> I was asking for examples of something that MS has
>> done that can't be done by following cross-platform
>> standards. Please give some examples of this 'progress' that
>> can only be done at the expense of trapping the user into
>> a single platform.
>
>Sure. OLE. Do *that* with some cross-platform standard.
You mean release a dozen non-interoperating flavors in
the span of a couple of years? Cross-platform standards
have no reason to make it difficult for competitors to
follow. Or do you mean something obscure that only
works on the local machine? Why bother when there is
corba and rpc which don't care about the location of
the target?
>[snip]
>> >I think part of your argument is missing. This doesn't
>> >seem to lead anywhere as it stands.
>>
>> Sell, give away, embed tools that claim to follow standards
>> but in fact only work correctly with other MS products. I
>> don't have to make this argument - the Halloween document
>> explains the philosophy.
>
>Sure does- but you aren't really giving it a fair shake. It's not
>that MS products won't work 'correctly' with non MS products;
>it's that they get more features when they work with
>other MS products.
No, it is that they don't work correctly with standards-conforming
competing products.
>[snip]
>> Which Windows mailer lets you easily pipe individual messages to
>> your choice of other programs?
>
>None that I know of, I'm afraid: I did not realise at first
>that you meant *mailers*; still if it were useful it could
>be done- Windows does have pipes.
Virtually all do on unix, hence the simplicity of adding
capabilities without having to change anything.
>[snip]
>> >Oh course. But those programs must know the protocol;
>> >I don't think you've bought yourself anything here.
>>
>> Yes, for an incremental change: you just add an extra content-processing
>> program in the places that need it without distrupting anything
>> else, including the way people have been reading their mail.
>> At his convenience, each user can switch to a new mailer
>> if he wants, and is ready.
>
>There are, of course, reasons why many people prefer Windows;
>this is one of them.
I think you got that backwards - changing any of the non-standard
mailers on Windows requires a complete all-or-nothing cutover.
>[snip]
>> >Really. How do you deal with CR/LF translation then?
>>
>> The transport already takes care of the local text conversions,
>> even to non-ascii platforms. That's the point of cross-platform
>> protocols.
>
>And you're going to be able to read a text file attachment
>even though it has been encoded so that these translations
>won't happen?
Yes, unsurprisingly the standards bodies anticipated this
well understood issue and the content-type of each attachment
is noted for correct handling.
>[snip]
>> >Well, unless you have a program that can cope with
>> >that format. Proprietary or not, makes no difference.
>>
>> Of course it makes a difference. If you know the format
>> you can write the program,
>
>This is not realistic for most people.
Fortunately, only one person really has to do it.
>> unless someone else beats
>> you to it. If it is kept secret you are forced to wait
>> until someone wants to sell you the ability to read
>> your own data.
>
>You'd be surprised how many people like to sell
>that, actually. :D
I'm not surprised at all. What surprises me is the number
of people who continue to be willing to pay, now that
it generally isn't necessary.
>[snip]
>> >I think it's possible to be too picky about that;
>> >Calling IE's language something other than
>> >"HTML" is going to obfuscate more than it clarifies,
>> >in my view.
>>
>> No, the correct way would be to use the version number
>> of the specification that it completely supports.
>
>I think that having a generic term "HTML", no version number,
>is useful and desirable.
It would be useful and desirable if the generic term meant
exact compliance to the current version of the standard.
Otherwise it is mainly useful for deception.
>[snip]
>> >Why does "follows standards" buy you anything
>> >that way?
>>
>> The cross-platform standards assure you that you
>> can change any component without modifying the
>> others.
>
>I think you know why I do not believe this.
No I don't. I assume, since you always advocate
anything that allows MS to gouge more money from
their customers, that you have some sort of vested
interest in Microsoft. But maybe you just don't have any
experience in actually using better platforms.
>[snip]
>> Except J++, Frontpage, anything that uses COM, Active Directory,
>> MS-CHAP, and the list goes on.
>
>Your exampes are mostly false, if not entirely so. COM, for instance,
>is available from several sources. J++ can work with
>non-MS JVMs and MS's JVM an work with non-J++ stuff. Etc.
>
>The only one I'm not sure of is MS-CHAP. Did that ever
>get documented and implemented by others?
Yes, the wide deployment made it necessary for others
to become equally insecure. The differences from
RFC 1994 CHAP are noted here:
http://www.cisco.com/univercd/cc/td/doc/product/software/ios113ed/113t/113t_3/mschap.htm#xtocid85670
among other places. I think the extension was submitted
as an rfc that may have been accepted by now, not that it
matters any more since following standards obviously isn't
important when Microsoft feels like changing something
to make interoperation with competitors difficult or impossible.
[snip]
>> Where's that non-MS Active Directory server?
>
>Active Directory is a trademark of MS, I'm sure, and
>the'll sue your butt of if you call *your* network
>server that.
Neat - yet another thing designed to take away the
choices of their users.
>But there *are* other network directory servers
>out there, ya know.
None are supplying clients tied to a monopoly
operating system.
>[snip]
>> >Surely. But I'm also 'allowing' that vendor to support
>> >*multiple* protocols- like that PCI-and-ISA computer.
>>
>> Also? Why would anything ever limit this ability?
>
>You would- it's not *standard*.
Huh? Following multiple standards is just fine and I have
never said otherwise. In fact I gave a specific example of
allowing email access through both POP and IMAP.
>> It is doing the standard ones correctly that is important,
>> though, as opposed to propagating vendor-specific extensions.
>
>I don't agree. The standard ones are important if and only if
>the computers you want to interoperate with are using
>them.
Yes, but extend that to the computers you want to interoperate
with tomorrow. If you use non-standard protocols or worse,
something that deceptively uses the name of a standard but
in fact does not follow it, you will be locked into that
particular vendor. You may have your reasons for wanting
everyone to be locked into using MS products forever, but it is
not in the user's interest.
>[snip]
>> >> Buy some non-Intel compatible CPUs and see how well
>> >> Microsoft's solution works.
>> >
>> >Ah, you cut me to the quick! Well, as quick as I ever
>> >get, anyway. :(
>>
>> Which part of cross-platform don't you understand?
>
>The part where only "standard compliant" systems
>count.
So you really are locked into the MS/Intel platform?
[snip]
>> MS hardly
>> invented the concept. They just gave everything it's
>> own bizarre API so the programs you write can't easily
>> be ported to other platforms.
>
>Realistically, portability between platforms is *never*
>easy, unless you consider different variants of the
>same platform "different", like Windows 98 and NT.
I consider different variants of unix to be different.
Considering that they have nothing in common but
source compatibility (different vendors, different
code bases, wildly different CPU types) I don't
see how anyone can say they aren't different. Yet
I have ported code across several versions spanning
15 years and a bunch of CPU types with no trouble
at all.
>[snip]
>> >Conforming browsers are sometimes quite unable
>> >to display these pages.
>>
>> I don't see how needing to have a standard component that
>> would have a fail-over substitute mapping relates in any way
>> to needing extensions that are only available from a
>> single vendor on a certain platform.
>
>You are saying it's wrong to put up a page that a standards
>compliant browser cannot correct display.
I am saying that claiming something is HTML when it in
fact is not is deceptive and wrong.
>This limits the internet to the commonly available
>ASCII character set, or something like it.
No it doesn't.
>"Fail-over substitute mapping" won't do; there *is* no
>reasonable, general way to substitute Chinese
>characters into the Roman alphabet short of translation
>(and even that's sometimes kinda dubious)
The way to specify the character set is included in
the standard, so it is possible for a conforming browser
to display it correctly. If you have some point regarding
standards here, I fail to see it.
>[snip]
>> >Perhaps your notion of "exact" is not as severe as
>> >mine.
>> >
>> >Are "informed consumers" expected to know the ins-and-outs
>> >of protocols and formats?
>>
>> They have to now, or they are deceived into using vendor-specific
>
>So they *do* have to know the details of protocols, in order
>to be considered "informed" by you.
Yes, if a protocol is followed correctly you do not need to
know the details. If you use or try to interoperate with MS
products you end up having to know all sorts of ugly details.
>> extensions that will not work with competing products. I'd
>> prefer that they did not need to know this and that they
>> not fall into the vendor's trap.
>
>Well, the world isn't going to conform to your desires on this
>one; consumers will never be "informed" as you understand
>it. They don't wanna be, and if its Microsoft that lets them
>avoid it, then it's Microsoft they'll patronize.
No, it is Microsoft's differences from the standards that
make everyone have to know about them. There are lots
of people who are deceived of course, but it is only
a matter of time until they are exposed to the truth.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Micro$oft retests TPC benchmark
Date: Mon, 31 Jul 2000 15:21:32 GMT
In article <[EMAIL PROTECTED]>,
Jen <[EMAIL PROTECTED]> wrote:
> "We are particularly interested in Microsoft's COMWare architecture,
> centered around COM+, because we believe it offers companies the
> opportunity to build high throughput (100,000,000+ transaction per
> day) web based commerce systems with extremely low cost per
> transactions."
>
> Strange to see Fungus posting references to a pro COM+ site.
>
Strange thing that your quotation does not appear on the referenced
page! Where did you find it? I guess it is somewhere on the site, but
not in the article (issue 27) anyway.
Regards,
Fredrik
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Roberto Alsina <[EMAIL PROTECTED]>
Crossposted-To:
alt.fan.rush-limbaugh,misc.legal,talk.politics.misc,alt.politics.libertarian,talk.politics.libertarian,alt.society.liberalism
Subject: Re: Aaron Kulkis -- USELESS Idiot -- And His "Enemies" -was- Another
Date: Mon, 31 Jul 2000 12:36:34 -0300
"Aaron R. Kulkis" escribió:
>
> Loren Petrich wrote:
> >
> > In article <[EMAIL PROTECTED]>,
> > Aaron R. Kulkis <[EMAIL PROTECTED]> wrote:
> > >Roberto Alsina wrote:
> >
> > >> And what happens if you die? Should your mother be killed,
> > >> or left to die by her own means?
> > >There's always my brother, and if he dies, my cousins.
> >
> > However, if Mr. Kulkis gets into a car accident far away from his
> > brother, and everybody else starts acting in Kulkis fashion, then the
> > last thing he will ever experience will be people around him snickering
> > good riddance to a total loser and how we must rid the gene pool of those
> > who get into car accidents.
>
> Hey, that's life.
Fuck no, that ain't a life. Do you live in some sort of Escape from New
York
nihilistic fantasy?
--
Roberto Alsina
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: The Failure of the USS Yorktown
Date: 31 Jul 2000 15:22:27 GMT
In article <[EMAIL PROTECTED]>,
Colin R. Day <[EMAIL PROTECTED]> wrote:
> Jim Richardson wrote:
>> The Sherman was cheap, but that's about it, (oh, and the
>> hydraulically stabilized turret was better, it could shoot
>> reasonably accurately whilst moving.) But it had a myriad of
>> flaws. It used radial aircraft engines, they used gasoline,
>
> The US was the world's largest oil producer at the time, so why is this
> a problem?
Gasoline engines tend not to be so efficient, and gas is only useful
to a tank when it is in the tank's tank. So to speak. That makes you
more dependant on your supply lines. Which is dangerous in war...
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- I may seem more arrogant, but I think that's just because you didn't
realize how arrogant I was before. :^)
-- Jeffrey Hobbs <[EMAIL PROTECTED]>
------------------------------
From: Roberto Alsina <[EMAIL PROTECTED]>
Crossposted-To:
misc.legal,talk.politics.misc,alt.politics.libertarian,talk.politics.libertarian
Subject: Re: Anti-Human Libertarians Oppose Microsoft Antitrust Action
Date: Mon, 31 Jul 2000 12:41:49 -0300
Stuart Dunn escribió:
>
> > > The robber barons want H1-B visa employees at slave wages...
> > ^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^
> You sound like a socialist. Employers, by definition, do not enslave
> their employees. The foreigners who are coming over by using H1-B visas
> are coming here because they have more freedom here than in their home
> countries.
Having several friends who did just that, each one said the same: "I am
doing it for the money". Noone ever mentioned freedom. Then again, the
US
is not really much more free than this country.
What the heck, you can drink beer on the streets, here.
--
Roberto Alsina
------------------------------
From: "Mike Byrns" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Micro$oft retests TPC benchmark
Date: Mon, 31 Jul 2000 10:38:12 -0500
"fungus" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
>
>
> Mike Byrns wrote:
> >
> > Everything is redundant.
> >
>
> No it *isn't*, that's the entire point. How many times must
> it be pointed out to you?
You haven't proven that assertion, ever. Show a non-advacacy site that
offers evidence of that fact and I'll shut up. What is comes down to is --
that's your best excuse.
> The Microsoft TPC database is split into equal chunks, if
> any chunk/machine is lost then all processing stops for that
> part of the database.
Nope. The data is redundantly distributed. SQL Server 2000's transactional
replication provides for loose consistency between a publisher and
subscriber when your application demands not just identical data at
different sites, but the necessity to mirror each and every data update,
addition or deletion, so you can reliably track every change to your
publication in close to real-time. The data is replicated on all machines.
It can be worked with in parallel with the system handling the transactional
replication behind your back.
> If part of a big (eg.) Sun server fails all parts of the
> database keep on running.
If one of the servers in the MS cluster fails everything keeps on running
too. Since everything is replicated.
> If you really can't see the difference then you're beyond
> hope.
There is no difference besides the fact that Oracle and nix can't do this.
------------------------------
** 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
******************************
