See also: last week's Back page page.

Linux links of the week

Section Editor: Jon Corbet

Letters to the editor

Date: Wed, 09 Feb 2000 23:45:56 +0000
From: kevin lyda <kevin@suberic.net>
To: letters@lwn.net
Subject: Your RedHat news item.

You stated it had "hype" in it and quoted the comment in the mail that
it eats "Lizards for breakfast."

You know what sucks about RedHat's IPO?  They can't make jokes without
people getting all uptight.  The piglet announcement also listed "It
works" as a new feature for it's rescue disk.  Was that anti-hype?  And
usage number 4: fodder for a thesis on "works in progress"; that seems
like hype?

Let them have fun fer cryin out loud!  If you'll note, their installer
is called Anaconda which probably eats lizards so it's an old joke (or
old hype) that was too subtle I suppose.

Linux is supposed to be fun.  The RedHat job page used to list "Nerf
guns" as sign-on bonuses, dunno if they still do, but I'm guessing they
still like to have fun.  It's nice to see it spill over into their
announcements, but with "hype" complaints it's easy to see why more
established companies buckle under to a more sterile outward appearance.

Please lighten up.  A little ribbing between rivals is a nice human
hackery type of thing to see.

Kevin
-- 
kevin@suberic.net                              Nutrition Facts
fork()'ed on 37058400		       Puns: 100% RDA  (% good puns: 0)

Date: Sun, 06 Feb 2000 13:09:36 -0500
To: letters@lwn.net
From: Bruno Majewski <bruno@pubnix.qc.ca>
Subject: About the inclusion of JFS into kernel 2.4

Hello --

I will admit to have been completely surprised when I read that IBM not 
only ported JFS to Linux (or at least is working at porting it right now), 
but even GPL'd it!  It looks like Big Blue is becoming a company  you can 
like and respect, instead of being the company everyone likes to hate -- 
quite a change, an almost unbelievable one for the generation that saw the 
original cassette-based IBM PC (yes, Virginia, there was life before the 
omnipresent hard disk).  I just hope this is not just a dream...

A JFS is, with raw I/O and other things like 32bit UIDs, just what Linux 
needs to become even more respectable in the corporate heavy-iron spheres. 
And it would be very nice if it could be incorporated into the kernel ASAP 
-- like, in kernel 2.4 if possible. Since 2.4 is not going to be out in the 
next few weeks, if we will have to wait a bit longer anyway, could it be 
possible for the IBM crew to put the pedal to the metal and work their 
collective butts off to finish porting / integrating JFS into the 2.3.x 
series of kernel so that 2.4 can come out with a JFS?

I already consider kernel 2.4 to be a major upgrade for the Linux world 
when it comes out to the point where I wonder if it should not be called 
3.0. I believe that the 2.4 kernel and XFree86 4.0 combination will be a 
major milestone in OSS evolution and for the Linux marketplace.

BUT we also have to consider against what it will have to go against in the 
marketplace: Windows 2000.

There are enough PHB and "Unknowing Deciders" (those who control the 
budgets) that are already truly excited about W2K that Linux has to gain 
every bit of functionality it can to compare favorably to M$'s offering. 
This is why I believe that, if it is possible, JFS should be rolled in 
2.3.x ASAP to that 2.4 can come out with it. As we currently stand, Linux 
is not quite ready yet for the desktop, but already has a reputation in the 
server area -- and if it wants to compare well to W2K as a server, the 
inclusion of JFS can only help (JFS on a desktop could be overkill). Y2K 
could become The Year Linux Became An Unavoidable Market Player (TYLBAUM 
?), the year where it can walk through the front door of more than just the 
oddball company, if it has all the right bits in place. And one of these 
bits is a JFS.

If W2K was only coming out in 3QY2K, I could live with a 2.4 kernel without 
a JFS. But W2K is coming out _this month_ and too many companies have been 
playing with betas of it for the last few months. So right now, too many 
managers and other "professionals" are enamored with W2K while still 
dismissing Linux as a toy. This is why we need to make 2.4 _the_ Killer 
Kernel, the base for a Killer Linux Distro -- and JFS can be a good way to 
attain that status.

So, can it be done? Can IBM finish porting JFS to Linux in time for it to 
make it into kernel 2.4? Can the IBMers can coordinate with Linus and Allan 
so that we could we see JFS-equipped 2.4-based distros this summer?

Bruno Majewski
bruno@pubnix.qc.ca

Date: Fri, 04 Feb 2000 08:32:16 +0000
From: Mike Goldman <whig@by.net>
To: letters@lwn.net
Subject: "Flame" sent to LinuxOne

While many distributions began as derivatives of others, something not
only allowed but encouraged in the Free Software / Open Source
community, such "forks" almost always provided some enhancement not
otherwise available, and returned something to the community.  LinuxOne
seeks to share in the "spoils" of the great commercial valuation which
many GNU/Linux vendors have recently enjoyed, without first even
creating a differentiated distribution or service of their own.

In this, LinuxOne is not only behaving unethically, but does real damage
to the credibility of other firms who have invested substantially in the
improvement of free software.  LinuxOne's alleged distribution provides
nothing but what is already available in RedHat and Mandrake offerings,
and they have no demonstrated capability even to provide support for
what they are selling.

Assuming that LinuxOne even manages to hold an IPO, presuming that their
sole underwriter doesn't abandon them, and that Linus Torvalds even
permits them to use "Linux" in their company name, both of which are
becoming increasingly dubious propositions, it is likely that any shares
which unsuspecting investors purchase would quickly collapse in value to
a marginal price.  Given that much of the investment media has no idea
how to differentiate LinuxOne from other GNU/Linux firms, this will
almost certainly result in a great deal of bad press for the whole
community, causing a stock devaluation for other firms who have little
or nothing in common with them.  This is unfortunate, and avoidable, if
LinuxOne were to voluntarily withdraw their offering indefinitely, or at
least until they have earned some credibility.

Should you decide to go forward with your IPO notwithstanding the
unnecessary harm you would inflict by doing so, any profit you might
make by doing so are tantamount to fraud perpetrated against those who
do purchase your shares.

Date: Thu, 3 Feb 2000 13:17:26 -0600
From: "John J. Adelsberger III" <jja@wallace.lusArs.net>
To: letters@lwn.net
Subject: Feb. 3rd OpenSSH 1.2.2 announcement

I'm glad you included this item, because I believe that OpenSSH will
eventually be the only way to fly on Unix-like systems.  However, I
would like to point out that the quoted text is somewhat misleading.

The OpenBSD team created OpenSSH, and one of the things they did was
to rip out all support for other operating systems.  This made the
code easier to audit because it is relatively small.  The people who 
port OpenSSH to other platforms have the benefit of this audit work,
but it is unfair to the OpenBSD team to say that the resulting ports
have been audited.

This is especially true given that in the past, security products ported
to Linux have quite frequently been found to have problems caused by
inadequate care taken by people doing the port.  It is not reasonable
to put the OpenBSD team's reputation on the line for such a product;
they wouldn't willingly do it, I'm reasonably sure(though I don't speak
for them,) and nobody else should either.

Do understand that I have nothing against the people who did this work.
In fact, I know nothing about them.  They may be, and probably are,
good honest people making their best effort - and the result may well
be good solid code.  It is not proper, however, to stake someone else's
reputation on that code.

-- 
John J. Adelsberger III				ETAONRISHDLFCMUGPYWBVKXJQZ
jja@lusars.net