Linux-Advocacy Digest #474, Volume #32 Sun, 25 Feb 01 20:13:03 EST
Contents:
Re: .NET is plain .NUTS ("Edward Rosten")
Re: Why don't we see more advocacy for Linux/MPI? ("Edward Rosten")
Re: .NET is plain .NUTS (Bloody Viking)
Re: RTFM at M$ (T. Max Devlin)
Re: RTFM at M$ (T. Max Devlin)
Re: Microsoft says Linux threatens innovation ("Erik Funkenbusch")
Re: .NET is plain .NUTS (mlw)
Re: Microsoft says Linux threatens innovation ("Erik Funkenbusch")
Re: RTFM at M$ (T. Max Devlin)
----------------------------------------------------------------------------
From: "Edward Rosten" <[EMAIL PROTECTED]>
Subject: Re: .NET is plain .NUTS
Date: Mon, 26 Feb 2001 00:08:19 +0000
> Why people use warez is becuse
> very likely they like the software but are not going to pay for it
> anyways. Office is a good example. It could be nice software, but I
> sure would not pay $300 for it. I would just as soon use a fucking text
> editor. And sure as shit, that is exactly what I do. I use a text
> editor for my "word processor"
Get TeX/LaTeX as well, because editor+TeX = far better quality output
than MS could ever *dream* of and at a fraction of the effort required to
write a long document in a wordprocessor.
-Ed
--
| u98ejr
| @
Share, and enjoy. | eng.ox
| .ac.uk
------------------------------
From: "Edward Rosten" <[EMAIL PROTECTED]>
Subject: Re: Why don't we see more advocacy for Linux/MPI?
Date: Mon, 26 Feb 2001 00:03:13 +0000
In article <[EMAIL PROTECTED]>, "Nic Bellamy"
<[EMAIL PROTECTED]> wrote:
> mmnnoo wrote:
>>
>> If you wrote a distributed encoder for mp3 and movie clips, or perhaps
>> a good distributed compilation system, you might start to see some
>> college dorms transformed into supercomputing centers.
>
> Have you seen the MPI version of Bladeenc? It doesn't seem to be
> maintained anymore, but may be of interest -
> http://www.mpi.nd.edu/~jsquyres/bladeenc/
>
> MP3 encoding is not actually a very fun thing to distribute - it keeps
> some context from frame-to-frame, so you need to either (a) share the
> context, or (b) put up with the fact that an mp3 encoded via MPI will
> not be identical to one encoded in a single run by one encoder.
It is generally easier to split MP3 encoding if you split it up by track
(since there isn't usally any context carried from one track to the
next). Then you just put a non distributed encoder on each machine and do
each track seperately.
-Ed
--
| u98ejr
| @
Share, and enjoy. | eng.ox
| .ac.uk
------------------------------
From: [EMAIL PROTECTED] (Bloody Viking)
Subject: Re: .NET is plain .NUTS
Date: 26 Feb 2001 00:20:31 GMT
Andres Soolo ([EMAIL PROTECTED]) wrote:
: Well, I know it isn't gnu.misc.discuss, but I agree with RMS on that point.
: Helping friends is more important than paying for some third party.
: Even if made unlawful by this third party, it doesn't mean its unethical.
That is an evil of payware. The copyright thing makes it illegal to to help
out your fellow human being. A perfect example of this evil is a techie giving
a techie wannabe an album with a supercostly server OS like NT Server. If the
techie obeys the law, the would-be techie is deprived of a lot of potential
income. If the techie breaks the law, a new techie emerges and that person
finds a way out of poverty.
--
FOOD FOR THOUGHT: 100 calories are used up in the course of a mile run.
The USDA guidelines for dietary fibre is equal to one ounce of sawdust.
The liver makes the vast majority of the cholesterol in your bloodstream.
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: RTFM at M$
Reply-To: [EMAIL PROTECTED]
Date: Mon, 26 Feb 2001 00:31:35 GMT
Said Erik Funkenbusch in alt.destroy.microsoft on Sun, 25 Feb 2001
>"Mart van de Wege" <[EMAIL PROTECTED]> wrote in message
[...]
>Fine. The host *DOES* implement ICMP echo.
Apparently, it doesn't. I sent a ping to www.microsoft.com, and
received no response.
>The router inbetween the host
>and the outside world doesn't allow it to pass. I know of no RFC which
>requires a router to pass ICMP traffic.
You don't seem to understand the workings of TCP/IP. What care I about
a router? If there is no response, there is no response. The only
possible explanation is a failure somewhere.
--
T. Max Devlin
*** The best way to convince another is
to state your case moderately and
accurately. - Benjamin Franklin ***
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: RTFM at M$
Reply-To: [EMAIL PROTECTED]
Date: Mon, 26 Feb 2001 00:31:36 GMT
Said Erik Funkenbusch in alt.destroy.microsoft on Sun, 25 Feb 2001
[...]
>And MS's implementation *DOES* support it. It's router just doesn't allow
>it inside it's network.
What's with this "its just" stuff? If I wanted to have to know how MS
implements Internet connectivity, I'd have asked. All I need to know is
that it is broken. Broke-dick, to use the technical term.
--
T. Max Devlin
*** The best way to convince another is
to state your case moderately and
accurately. - Benjamin Franklin ***
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: Microsoft says Linux threatens innovation
Date: Sun, 25 Feb 2001 18:37:09 -0600
"Ed Allen" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> >I don't care if the pope contradicts me. These are hard cold facts.
> >
> >A) It's a fact that the number of dollars charged for Windows has not
> >changed in 6 years.
> >B) It's a fact that inflation has increased in those 6 years.
> >C) It's a fact that we now recieve more software with windows than we
did 6
> >years ago.
> >D) It's a fact that when inflation increases, the value of the dollar
> >decreases.
> >
> >E) It's a fact that if number of dollars stay the same while inflation
goes
> >up, that the actual cost of the product has gone down.
> >
> >No amount of email evidence you try to put up will change these
economical
> >facts. Stop pretending it's not true.
> >
> Both you and M$ will find that those Findings of Fact cannot be
> ignored forever.
I'm not ignoring the FoF. They are simply irrelvant to this portion of the
argument. The sun sets in the west may be a fact, but it's irrelevant to
price of noodles.
> Jim Allchin is the M$ executive who was in charge of IE during the
> time in question.
>
> He said that the OEM price was raised, not the retail one. You have
> no way of knowing what the OEMs paid. Those numbers were redacted
> from the court records and the contracts themselves remain secret.
What are you talking about specifically here? OEM's pricing was subject to
volume and several other factors. If an OEM's volume goes down, it's price
can go up. It's MSRP has never changed or been exceeded.
> If the retail price of the Windows bundle was ten times what an OEM
> paid in 1989 and the same number of now inflated dollars were only
> three times what that OEM paid in 1995 then Jim Allchin would be
> right that the price went up and it would appear to you that it had
> not.
OEM's didn't buy Windows in 1989. And even if they did, it was a completely
different product then. I worked for a major OEM in 1993 (Zeos), and I know
what they paid for Windows and DOS (I maintained their MRP system). The
price of Windows 95 was much larger than the price of Windows 3.1, but then
Windows 95 was a standalone product, while Windows 3.1 required the seperate
purchase of DOS.
> >> I have never claimed that MS makes innovation totally stop just
that
> >> it raises the cost of bringing such innovation to market.
> >
> >Then explain how Debian is able to bring them to market for free.
> >
> Because Debian is depending on contributions from around the world
> to support the users in getting it installed and fixing their
> problems.
That still doesn't explain how if the cost of bringing things to market has
raised because of MS, that Debian is still able to do so for free. Whether
or not people donate time, it's still free for Debian, thus the cost has not
risen for them to bring their product to market.
> >> That is finally beginning to allow people to see that not all
> >> computing needs to be under the thumb of MS.
> >
> >Then what's your argument?
> >
> That M$ has been deluding millions of people into thinking that only
> Bill can have good ideas.
That's not criminal, whether or not it's true or not.
> That M$ has been using the preload lockins to exclude any non Bill
> approved ideas from the desktop.
Not true either. They have had limitations on the "first boot", but OEM's
are allowed to put anything they want on the desktop, and they always have
been. They just haven't been allowed to remove things from the desktop.
> That ideas are permitted to open new markets outside the desktop so
> that Bill can study them away from the view of the faithful and then
> to buy or steal them to be bolted on to the Windows bundle if their
> new market looks profitable enough.
"permitted". You act as if MS controls everyone. They don't. Otherwise
Java would not exist. Linux would not exist.
> >Many markets have large financial barriers to entry. You can't come into
> >the automobile market without spending billions in factory and R&D costs.
> >You can't get into the processor market without spending billions. You
> >can't provide long distance service without spending billions in
> >infrastructure. That's just the cost of getting into a very well
> >established market.
> >
> Hardly something we can reasonably expect to come from a "garage"
> operation though is it ?
Yet it's happening, isn't it?
> Shows the M$ lie of "We can be put out of business in five years
> by two kids in a garage" for how false it truly is.
Where do you get this "two kids in a garage" quote from?
> >> As for "Linux's technology has not advanced past the level that
MS's
> >> has". Since Linux compiles the same kernel to run on the iPAQ and
> >> the z900 and MS require different source trees for servers and
> >> desktops on x86 alone how can you justify saying that ?
> >
> >NT/2000 uses the same source tree for Server and Desktop.
> >
> But they require WinCE for embedded. Linux does not require a
> separate tree for even multiple different CPUs.
I guess it depends on what you mean by same source tree. Embedded linux
most certainly has it's own unique branches and it lacks much of the
functionality of a full fledged desktop or server build.
> Linux uses the same source tree in the Beowulf clusters among the
> Top 500 supercomputers.
>
> Is Datacenter the same as Desktop ? If it is then what makes it
> worth several thousand dollars per CPU ?
Datacenter comes with all kinds of extra software, while the basic kernel is
the same between all builds.
> >In any event, we're talking about end-user technology, not OS developer
> >technology.
> >
> I notice that M$ began to bring out "skins" for XP but is now
> holding them back.
No, it's going to be skinnable. It's just that the skinning sdk won't be
publicly available. This is to keep the user interface consistent.
> Linux has had "skins" for at least three years.
And you've been able to use products like WindowBlinds for years as well.
> Who is ahead again ?
>
> As for developers being left out, one of the primary observations of
> the "Halloween Documents" was that any user could become a developer
> because the development tools came on the CD.
>
> That is another "Barrier to Entry" being reduced by Open Source as a
> whole, not just by Linux.
The Halloween Documents were not an official public statement. It was the
opinions of one technician.
> >> Linux has grown from nothing but a mutitasking plaything on Linus'
> >> desk to running several, I have pointers to ten at least, computers
> >> in the top 500 largest supercomputers in the world. I have a
single
> >> pointer to a W2K member of that group. How is Linux behind by that
> >> measure ?
> >
> >One data item is not a complete picture. Linux has nothing like DirectX.
> >Windows has more features that beat Linux's features than vice versa.
> >
> No Windows has proprietary features denied to Open Source by NDA and
> contract exclusions.
Sorry, but directly accessing hardware through an abstraction layer is not
something "denied" to Open Source.
> That Linux does not have lots of Win32 functions is due to hard work
> and money buy M$.
I said "like" DirectX. It doesn't have to have the exact API's. It should
just have something equivelant.
> If your measure of superior is how many Win32 functions are
> supported then Windows will always be "better."
>
> Hardly seems an objective conclusion though.
DirectX is the primary reason that game developers switched from DOS to
Windows. Clearly this is something that game developers found highly
useful.
> >> Linux today, in delivered code, functions perfectly on 64 bit CPUs.
> >
> >Great. MS has shipping versions of NT that also run on 64 bit CPU's. NT
4
> >runs fine on Alpha's, which are all 64 bit.
> >
> That "shipping code", which is no longer supported, is only able to
> run in 32 bit mode though.
That's not what you said. You claimed that Linux functions perfectly on 64
bit CPU's, insinuating that Windows didn't function perfectly on any 64 bit
CPU's. It did (and still does, even if it's not supported).
> M$ was working on a 64 bit version of NT and had some falling out
> with CompaQ so the, formerly DEC, engineers were directed to stop
> work till it was cleared up at which point M$ terminated the
> agreement.
No, Compaq terminated the agreement. They didn't want to spend money on NT
Alpha support since people weren't compiling their apps for Alpha.
> They may have had some alpha, not Alpha, code which they can still
> use internally but that code cannot be shipped.
Compaq and MS acknowledge that they were using the Alpha platform to develop
the 64 bit extensions to Windows prior to Itanium being available in
hardware.
> The only known 64 bit code runs on Itanium, not shipping yet, so M$
> provides it to developers who get a preproduction Itanium. Again
> not shipping yet.
>
> And Linux is still behind ?
Again, Windows has more features that Linux has no equivelant to than the
other way around.
> >NT can run on multiple architectures, simply because it has the ability
to
> >do so is the technology, not whether or not it actually is. NT has run
on
> >PPC, MIPS, Alpha, and x86 in shipping production code.
> >
> All withdrawn because M$ did not have a symbiotic relationship with
> the other CPU vendors.
No, because those vendors stopped selling machines which could run NT.
> Their relationship with Intel is falling apart, do you think any
> other company will cozy up to them like that again after the Intel
> testimony at the Antitrust trial ?
To claim that it's "Intel testimony" is a bit dishonest. It was one
employee that, despite upper managements pressure, testified. Word has it,
he was inches from losing his job because Intel upper management didn't
agree with his testimony. In fact, the Intel lawyers wouldn't allow the DOJ
to meet with him, and he didn't provide any written testimony prior to going
on the stand. The DOJ took a huge chance putting him up there, because they
literally had no idea what he was going to say.
> >> The WebBench figures that Chad was blathering about put W2K 2%
> >> behind Linux while running on faster hardware. How does that put
> >> Linux behind ?
> >
> >2% is not much of a statistical difference. Most tests have more than a
2%
> >margin of error. Show me even one commercial server running Tux.
> >
> Two percent behind on twice as many, 50% faster disks, seems like a
> pretty bad statistic to me.
Only if the speed of the disks was the bottleneck. I doubt it was.
> Show me the "even one commercial server running" whatever M$ is
> calling that WebCache thingy they say is not due until May 2001.
The 3.0 version isn't out yet. The 2.0 version is currently available.
> We all have great confidence that M$ never slips a product delivery.
Oh yes, like the the 2.4 kernel was on-time.
> >> Not taking defensive measures when you are being attacked is not
> >> healthy.
> >
> >Yet MS isn't supposed to do the same. Your hypocricy is showing.
> >
> M$ is the one who offered to divide up the market with the threat of
> economic death implied by refusal.
There is still a lot of doubt surrounding that meeting. I don't know if
we'll ever know for sure what really happened there.
> >Of course not. However, nowhere can the law be interpreted as saying
that
> >"acting like it's a life or death situation" is criminal.
> >
> Attempting to put competitors out of business is.
>
> Corporate "murder" is called "restraint of trade."
>
> Attempting it is called "monopolization."
>
> That is illegal, just as attempted murder is for you and me.
And none of those things are "acting like it's a life or death situation".
The OP (I think it was Max) stated that acting so was illegal. It's not.
> >A customer can only compare the value of a product to themselves. For
> >instance, a pickup truck can compete with an SUV. They both carry
people,
> >they just offer different sets of extended functionality and are priced
> >differently because of it.
> >
> Either they can replace each other or not.
Well, they certainly can replace each other for many customers.
> If the number of customers who find them interchangeable is large
> enough to be a significant share of the customers of each separately
> then they compete.
Yes, indeed they do. That doesn't mean their prices should be the same
though.
> If the overlap is only a small fraction of one who has an entrenched
> market which is not in the overlap then there is danger of
> monopolization.
I don't follow your point. We are talking about comparing prices here.
> So Erik,
>
> >> >> >> Are you denying that Microsoft has a monopoly ?
In the traditional sense, yes i'm denying that. In the legal sense, I think
it's a fine line. There are valid arguments either way.
> >> Diesel trains are more expensive than Dodge trucks and they both
> >> haul things. Are they competitors ?
> >
> >See above.
> >
> You did not answer my question. I did not ask about SUVs.
Stop being anal. The arguemnt is the same whether it's SUV's or
locomotives.
> >> I contend that you are mixing fruit again Erik.
> >>
> >> No other OS competes for the desktop. Can't, OEM preloads add the
> >> cost of whatever MS is calling the Windows desktop this quarter to
> >> any potential competitors' cost.
> >
> >Not true. As we've already seen, Linux is being loaded by OEM's without
the
> >cost of Windows. BeOS is offered on certain Hitachi systems, again
without
> >the cost of Windows. IBM *STILL* sells OS/2 on computers without the
cost
> >of windows.
> >
> The only competition between Linux and any M$ product is W2K with
> IIS, SQL Server, and Exchange included.
So, you're contention is that Linux cannot compete with MS on the desktop.
Glad to know you believe that Linux is so poor there.
> That is thousands of dollars on a single CPU box up to tens of
> thousands on a eight CPU powerhouse versus the under two hundred for
> Linux in whatever configuration.
>
> How is M$ not charging far more than the competition ?
The competition is not only Linux.
> >> How can OS/2 or BEos or whatever "compete" when somewhere between
> >> $40 and $90 gets sent to Redmond for every copy sold ?
> >
> >That's not been the case for over 6 years.
> >
> That is the case extant as long as we have 'cliff tier pricing.'
Consumers buy a complete system, the OS is included in that. This is
similar to people buying a TiVo. It has the OS included, and there is no
way for them to buy it without the OS. This is the way consumer electronics
work.
> Are you claiming that OEMs lied in their evidence produced at the
trial ?
Which evidence are you speaking of, specifically?
> >> Linux does not compete on desktops. You can order machines
> >> preloaded with Linux from Dell, Gateway, CompaQ, HP, and IBM at
> >> least, all servers not desktops.
> >
> >Also untrue. Can you even bother to look this stuff up?
> >
> >http://www.dell.com/us/en/bsd/topics/segtopic_linux_000_linux_center.htm
> >http://www.dell.com/us/en/bsd/topics/segtopic_linux_001_linux_center.htm
> >http://www.compaq.com/products/desktops/linux.html
> >http://www.linux.hp.com/systems/index.html
> >
> No I do not go survey vendors every few months just to see what they
> offer. When I was looking for a computer last year those links were
> not there.
It's been this way for at least a year. You are simply wrong on this.
> >> The only MS OS aimed at servers is W2K and it costs several hundred
> >> on up to thousands of dollars per CPU while Linux is less than two
> >> hundred even if you take the same CD and install it on dozens of
> >> eight CPU servers.
> >
> >Linux is an enigma, and not indicative of everything. A 17 processor
> >Solaris license costs $60,000, a 65 CPU license $360,000
> >
> I am not saying that Unix vendors, like Sun, are not going to need
> to reduce prices as well but we are concerned with M$ here not a
> vendor who produces a low volume, compared with PC volumes, product.
The issue here is that in the traditional OS market, which Linux is not a
member of because they don't have to pay for their developers, OS's cost
money. Lots of money. That's what the majority of MS's competition does.
Novell, Solaris, HPUX, DGUX, Irix, etc...
These companies cannot compete with a free OS in the same way. Sun has now
started giving away their OS for <8 processors, but that's misleading
because they've built the cost of the OS into their hardware. Solaris x86
is free, but it's almost useless since it supports so little hardware.
> When volumes go up efficiencies of scale reduce costs and, when you
> have competition, prices.
Not in a market like this. Increased competition in software will actually
cause prices to rise, since economies of scale allow more profit per unit
the more units you sell. The fewer units you sell, the more your per-unit
cost is, the less lenient you can be in pricing or R&D.
> Sun writes software, Solaris, designed for hardware which has low
> volumes. They need to recoup their costs of writing that software.
>
> That does not excuse M$ for maintaining monopoly prices in the PC
> arena.
MS also needs to recoup their costs for writing their software as well, and
they spend a lot more on R&D and advertising than Sun does.
> >> Sounds like a large price difference to me.
> >
> >Linux is not MS's only competition. MacOS X will cost $129 for an
upgrade.
> >
> That also sounds like a large difference to me if MacOS X is aimed
> at servers instead of just desktops.
No, that's MacOS Desktop. MacOS X server is $499 (which is a good deal, but
I think this will change after MacOS X desktop is introduced. MacOS X
Server has been out for almost a year, and is way behind the desktop).
> How does their having a low price support M$ not overcharging ?
That's a high price. MacOS X Desktop upgrade is $129, while MS's upgrade
price is $89.
> >> People "buy" it because avoiding it is almost impossible, not
> >> because they like monopoly crapware.
> >
> >Ask any 10 computer owners on the street what OS they want to use. What
do
> >you think most of them will say?
> >
> If they would buy it without being forced then why does M$ insist
> that remedies removing their preload restrictions would be "too
drastic" ?
Their preload restrictions were largely covering the areas of trade dress
(changing their startup sequence and what ships with the OS (ie removing
Internet Explorer), which could negatively iimpact MS's image in the
customers eye because of something the OEM did).
> And "taking away their Intellectual Property" ?
>
> More like removing the knife they had been putting to OEM throats is
> my view.
OEM's choose to use windows. There are quite a few OEM's that don't choose
to use windows at all.
------------------------------
From: mlw <[EMAIL PROTECTED]>
Subject: Re: .NET is plain .NUTS
Date: Sun, 25 Feb 2001 19:36:47 -0500
Andres Soolo wrote:
> Well, I know it isn't gnu.misc.discuss, but I agree with RMS on that point.
> Helping friends is more important than paying for some third party.
> Even if made unlawful by this third party, it doesn't mean its unethical.
I disagree. The creator of copyrighted material should have control over its
destiny. If a software maker does not want people to share their work, that is
their right. That being said, I think that most software is just plagiarism
anyway. So it should not matter.
The last original idea was probably from Knuth.
--
The majority of the stupid is invincible and guaranteed for all time.
The terror of their tyranny, however, is alleviated by their lack of
consistency.
-- Albert Einstein
========================
http://www.mohawksoft.com
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: Microsoft says Linux threatens innovation
Date: Sun, 25 Feb 2001 18:38:46 -0600
"Ed Allen" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> >>Software is not hardware, and it doesn't follow the same market trends
of
> >>hardware.
> >
> >Then why did you list only the most trivial hardware components above?
> >
> Could it be that M$ makes mice and keyboards ? I do not know about
> cases and power supplies but I suspect that those had reached
> commodity prices by many years of lots of radios and televisions
> requiring them.
MS doesn't have a monopoly on keyboards and mice. Logitech makes a
significant number of them, so does Cherry (In fact, Cherry makes more
keyboards than Logitech and MS combined).
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: RTFM at M$
Reply-To: [EMAIL PROTECTED]
Date: Mon, 26 Feb 2001 00:36:17 GMT
Said Bob Hauck in alt.destroy.microsoft on Sun, 25 Feb 2001 19:43:44
>On Sun, 25 Feb 2001 18:02:28 GMT, T. Max Devlin <[EMAIL PROTECTED]>
>wrote:
>>Said Bob Hauck in alt.destroy.microsoft on Sun, 25 Feb 2001 16:43:53
>>>On Sun, 25 Feb 2001 04:55:07 GMT, Norman D. Megill
>>><[EMAIL PROTECTED]> wrote:
>>>
>>>> I thought the most common DoS attacks were SYN floods. I've never
>>>> heard of a DoS attack with normal, short, non-broadcast pings
>>>
>>>Imagine if 10,000 people all started sending one ping/sec to the same
>>>site. Now imagine one guy planting a remote-control trojan like Back
>>>Orifice or trinoo on a few hundred systems and sending 100 pings/sec.
>>
>>That's the point, Bob. Notice that this is an imaginary example.
>>NOBODY has ever heard of a DoS attack with normal pings.
>
>If you say so Max. I could swear that I heard of this years ago, but it
>doesn't matter. It is a theoretical possibility, which is all I was
>trying to say. It is arguable whether blocking echo reply is "clueless"
>even though it is against RFC's. OTOH, I agree with you that blocking
>all ICMP is a bad idea, as that breaks things besides just
>troubleshooting.
Actually, it is only the ECHO_REQUEST/ECHO_RESPONSE (ping) datagrams
that are mandatory. All others are optional to begin with. While the
designers of TCP/IP may have considered MTU discovery a rich feature, it
has and always will be useless on the Internet.
>There _have_ been attacks based on pinging broadcast addresses using
>packets with forged source addresses. Blocking ICMP to the broadcast
>address is a sensible thing to do and doesn't break any functionality.
Actually, there is a great deal of value in supporting ping to
'broadcast addresses' (it doesn't really broadcast, by the way), such as
SunOS supports (or at least used to; they probably broke this in the
last couple versions.) "Blocking" such a thing would be a monumental
waste of time, as is all of this paranoid mucking around with IP. The
only real security that ever existed, and will ever be possible, is
login level (the session layer of the OSI, if you're keeping track).
The 'exploits' of the type you described are simply "packet pumps", and
it makes no difference whatsoever what kind of traffic is used. Such a
DOS attack will forever be impossible to pro-actively guard against, and
firewalling ping is just cutting off your nose to spite your face.
--
T. Max Devlin
*** The best way to convince another is
to state your case moderately and
accurately. - Benjamin Franklin ***
------------------------------
** 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 by posting to comp.os.linux.advocacy.
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
******************************
