Linux-Misc Digest #71, Volume #21 Sun, 18 Jul 99 13:13:08 EDT
Contents:
Re: Solving the 1024 cylinder LILO problem ("Neil Koozer")
Re: Shortcomings of Linux? (Anthony Ord)
IR port ("mindrace")
Re: Marx vs. Nozick (Stefaan A Eeckels)
Re: Slackware problem... (Andrei A. Dergatchev)
Re: Red Hat is Crap!! (Juergen Heinzl)
Re: Can someone recommend.... (Ilkka Ollakka)
Re: Solving the 1024 cylinder LILO problem (John Reiser)
Re: Wordperfect and Linuxmall Confusion! (Stephen Chadfield)
----------------------------------------------------------------------------
From: "Neil Koozer" <[EMAIL PROTECTED]>
Subject: Re: Solving the 1024 cylinder LILO problem
Date: Sun, 18 Jul 1999 09:01:01 -0700
>"K> >>This isn't a lilo problem, it's a BIOS problem.
>"K> >
>"K> >It's both.
>"K>
>"K>
>"K> Yes. It is a lilo problem inasmuch as there is no reason to write lilo
to
>"K> use the bios at all. (for IDE drives)
> ^^^^^^^^^^^^^^^
>
>This is the kicker. Some of us use SCSI boxes. SCSI controllers come
>in various shapes and sizes... ISA, PCI, SCSI2, SCSI3, wide, ultra wide,
>etc. Lilo uses the BIOS for a *good* reason -- it allows Lilo to work
>with a *variety* of boot devices. Unfortunately, the BIOS standard is
>broken and has this stupid 1023 cylinder limit...
Yah, I'm not that familiar with SCSI, however, the kernel driver does manage
to choose what code fragment to use for each type. The lilo map loader
could use the same method of choosing, but instead of running the code
fragment it would insert it into LILO (the first stage loader that goes into
the MBR). I don't know how many bytes it takes to read SCSI, though.
If it turns out that the MBR is not big enough, there are 62 more sectors on
that initial track going to waste, and I think any bios can address this
much if it can address the MBR.
Another advantage of avoiding the bios is that the size of the kernel would
be irrelevant, thus eliminating all those posts about "kernel too big".
.....and I have somewhat of a suspicion that the current size of the kernel
is not the largest that it will ever be.
Neil.
Neil.
------------------------------
From: [EMAIL PROTECTED] (Anthony Ord)
Crossposted-To: comp.sys.amiga.misc,comp.os.linux.advocacy
Subject: Re: Shortcomings of Linux?
Date: Sun, 18 Jul 1999 16:43:58 GMT
On Fri, 16 Jul 1999 17:58:20 GMT, [EMAIL PROTECTED] wrote:
>Is this guy right on the money regarding the mentioned shortcomings
>of Linux?
No he isn't "right on the money" in a number of areas.
>In article <7mkmun$knv$[EMAIL PROTECTED]>,
> Holger Kruse <[EMAIL PROTECTED]> wrote:
>> Ancipital <[EMAIL PROTECTED]> wrote:
>> > However, a lot of the other comments are
>> > based on 2.0x, and to say that it's "the worst in the business" is
>> > just cynical, mendacious FUD.
>>
>> That is your opinion.
>>
>> There are three major TCP/IP stack families: BSD, Microsoft
>> and Linux.
AFAIR Microsoft paid Berkeley to write the Winsock stack.
>>And yes, out of these three Linux has come up by
>> far worst in my two years of interoperability testing. It is
>> the only stack for which I ever had to put explicit workarounds
>> in my code.
>>
>> Initially bugs I reported to the Linux developers were
>> dismissed as "no bugs", but after the recent discussion
>> and flamewars renewed talks with Linux developers could
>> confirm at least one of the bugs, and I am happy to say
>> that it is fixed in some of the more recent Linux versions.
>>
>> In any case, direct bugs are not really the main issue. The
>> problem is that Amiga Inc. needs a TCP/IP kernel that is
>> very solid, well peer-reviewed, well extensible, widely used,
>> well-understood, professionally maintained etc. From my
>> point of view Linux simply does not fit into that category.
This is where the FUD really starts.
>> It is maintained by hobby programmers,
No it isn't. It is maintained by Alan Cox who works for Red
Hat. It is his *job* to maintain the networking code (along
with other things like the Video for Linux project.
>>which by itself is not
>> bad, of course, but IMHO makes it unsuitable for use in
>> multimedia convergence systems and vertical markets.
What does a TCPIP stack specifically have to do with
multimedia? Is this just one of those "let's put a hype word
in here because it looks k3wl"?
>>For those
>> kinds of applications Amiga Inc. needs solid support because of
>> the fast pace TCP/IP is changing these years. Just yesterday
>> IANA announced that IPv6 addresses are now being assigned to
>> regional Internet registries, i.e. the IPv6 transition WILL
>> begin, probably very soon -- at a time when some issues of
>> IPv6 are still not completely hashed out (e.g. multicast
>> routing in IPv6). Microsoft is pushing these things very
>> hard, and I don't doubt that they will succeed, as usual.
In terms of the Internet, or indeed anywhere where they
don't have numerical dominance, they usually fall flat on
their arse most times.
>> For years Microsoft has been stalling IETF and developing
>> its own protocols, different from official standards,
>> hoping that wide deployment of their proprietary protocols
>> will eventually throw out official standards and give Microsoft
>> a head start. Sometimes they succeeded, but on the Internet
>> more often than not they failed. Bill Gates almost completely
>> ignored the Internet initially, and tried to side-step it
>> later. That was probably his biggest mistake ever, and he
>> knows it and has changed his strategy since then.
>>
>> For the last few years Microsoft has been trying to push
>> their own ideas and strategies through IETF as official
>> standards (e.g. DHCP, a Microsoft-invented extension to
>> BootP).
No it isn't. DHCP was postulated by Ralph Droms of Bucknell
University in RFC1531 (superseded by RFC1541) during 1993.
Some of the ideas came from Sun's dickless workstations.
What was M$ doing with TCPIP in 1993?
>>Basically Microsoft develops something on their own,
>> pushes it as a (hopefully, from their point of view, very
>> complex) standard, that is difficult to implement, and then
>> gets a head start from the work they have already done by
>> being the first company with a working implementation.
>> Everyone else has to catch up.
And where they screw up, is that if it is popular, they try
to keep it a closed standard, (often adding even more
proprietary stuff) and if its unpopular they try to make it
an open standard. Of course it is unpopular for a reason, so
this strategy tends not to work.
>> Of course this only works if Microsoft actually succeeds
>> in making their ideas standards, and recently they have,
Examples?
>> because of their very close cooperation with IETF and other
>> committees. Those committees used to be more tilted towards
>> Sun, SGI, DEC etc. but recently Microsoft has become dominant
>> almost everywhere.
Because they are very rich.
>> There are very few groups and organizations that can still
>> withstand Microsoft, and most of those that do have one thing
>> in common: they all use BSD,
Rubbish. I'm not saying they all use Linux, but I am saying
that there are some non-BSD users among them.
>>either by extending it, or by
>> incorporating it in their own proprietary OS or software.
>> This means instead of requiring every company to implement
>> and test all new standards on its own, the work can be split
>> up among companies, the results are published,
But the results aren't always, simply because of commercial
advantage.
>>and the
>> knowledge can be shared. Plus the companies can use their
>> existing infrastructure for testing. This way several smaller
>> companies with proper financial backing can make up for the
>> size of Microsoft. KAME is one of the projects that works like
>> that: a group of companies that cooperate on extending BSD for
>> IPv6, IPsec (encryption etc.) and other things.
>>
>> Back to Linux: the main concern I have is that Linux does
>> not fit into these power struggles, and that there is
>> probably a good chance that Linux will simply be "swept
>> away" in a few years if Linux cannot keep up with the pace
>> of development.
Linux is starting to drive development in some areas. Look
for kvoicecontrol.
>>Considering how amateurish (not meant in a
>> degrading way)
Yes it is meant in a degrading way.
>>Linux development works, I doubt that it can
>> hold up against Microsoft for long.
When Linux started out, it was a couple of guys in Finland.
Now there are thousands of support people out there. If this
makes Linux weaker and more susceptible to M$, then that is
an inversion of the historical definition of strength.
>>Development of Internet
>> protocol implementations is quite expensive. It requires a
>> lot of travelling (to committees, for interoperability
>> testing etc.),
Do you think this guy has heard of the Internet? Travelling
to communicate? Travelling to use software? Please.
>> purchase of hardware, renting of high-speed
>> Internet lines etc. Not a big deal for large companies,
>> but difficult for small groups and loose communities without
>> much of a financial structure. That is why I do not write
>> most of the more complex Internet protocol implementations
>> myself, from scratch, but rather port them from BSD. That
>> requires less testing and is much cheaper.
>>
>> I give BSD-derived operating systems a much better chance
>> than Linux, mostly because of their financial support, in
>> particular from Japanese companies (Hitachi, Toshiba etc.).
Then why is BSD not growing at the rate Linux is if it has
"a much better chance"?
>> Also, don't overestimate the switching costs from Linux
>> to BSD and the power of the installed base. Linux users
>> are NOT locked into the Linux OS in any way yet. For DOS
>> and Windows lock-in and switching costs have been strong
>> arguments why people could not switch to a different OS,
>> but for Linux they are not:
That's right. It also holds true for *BSD. If you program
stuff correctly, you could switch back-and-forth every day.
>> Linux is Posix-based, with a small set of system calls,
>> compared to the extremely convoluted Windows. The Linux API
>> can easily be emulated by BSD operating systems. NetBSD
>> does, as a matter of fact, i.e. many "Linux applications"
>> will also work with NetBSD (at full speed).
And the converse is true also.
>>The reason
>> why many people today choose Linux over BSD is because of
>> the all-present Linux hype and because Linux is available at
>> every computer store
Yes, and the CDs are cheaper - at least for me.
>>and easier to install.
*BSD is harder to install?
>>QNX certainly
>> also has the potential of fully emulating Linux system
>> calls and running most Linux applications (on x86 at least).
>>
>> If, at some point in time, a new BSD-based OS is pushed by
>> a big company, and that OS is more advanced in its networking
>> support and supports the Linux API then I would not be
>> surprised if that OS took over and people started using their
>> former "Linux applications" with the new OS.
It depends on 2 things.
Cost.
Benefits.
>> Because of that I think it is unwise to concentrate so
>> much on the marketing factor. Linux and its installed
>> base are not nearly as strong as operating systems in
>> the past.
Eh? Like PrimeOS?
>>IMHO the technical qualities and the financial
>> backing and stability should be more important, and I still
>> maintain that I don't believe that Linux TCP/IP has the same
>> potential to become one of the leading TCP/IP stacks in the
>> next ten years as BSD and Microsoft have,
Except at the top of this rant he said "There are three
major TCP/IP stack families: BSD, Microsoft and Linux." So
he's contradicted himself.
>>mostly for historic
>> and organizational reasons. IMHO that is a very powerful
>> argument against using the Linux kernel for AmigaOS.
No. The reason why Linux can not be used in AmigaOS is that
you can't proprietise it because of the GPL. Any
improvements they make to it for commercial advantage would
have to be shared so everyone would benefit.
Concentrating on the TCPIP stack (which is a small part of
any OS) is just a red-herring.
>> --
>> Holger Kruse [EMAIL PROTECTED]
>> http://www.nordicglobal.com
>> NO COMMERCIAL SOLICITATION !
And that boys and girls was an example of how to write FUD
to obscure your motivations. I hope you are suitably
impressed and have learned from it.
Regards
Anthony
--
=========================================
| And when our worlds |
| They fall apart |
| When the walls come tumbling in |
| Though we may deserve it |
| It will be worth it - Depeche Mode |
=========================================
------------------------------
From: "mindrace" <[EMAIL PROTECTED]>
Subject: IR port
Date: Sun, 18 Jul 1999 11:29:38 -0500
hey, anyone heard of being able to add an IR port to an already running
system? i haven't seen them sold separately anywhere, i was just curious if
it was possible to add an IR port to an IR capable motherboard. if anyone
can tell me, i sure would appreciate it, as i have some practical
applications in mind for it. l8zzz...
-mindrace
------------------------------
From: [EMAIL PROTECTED] (Stefaan A Eeckels)
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.linux.advocacy,gnu.misc.discuss
Subject: Re: Marx vs. Nozick
Date: 18 Jul 1999 13:39:35 GMT
In article <7mse68$6$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Richard Kulisz) writes:
> Socialism is based on cooperation and democracy while capitalism
> is based on competition (ie, War) and dictatorship. You can't run
> the world on destruction alone but you sure as bloody hell *can*
> run it on construction alone! The same applies to honesty vs. lies
> in moral philosophy. The situation is *not* symmetric.
>
> So while it's obvious that Libertarians are full of shit and idiots
> besides, the Marxists have hit upon a fundamental principle of nature.
Harumph. The natural world *does* run on competition, rather
than on cooperation: competition for food and living space,
both inter and intra species.
When left alone, nature weeds out the weaker, and the stronger
get to procreate.
Humankind in the last few centuries is an a-typical species, as
it seems to evolve towards favouring the weaker elements, and
has reduced the procreative advantage of the stronger elements.
Getting people to cooperate is *difficult*, because by nature we're
competitive. Humans need to be compelled to cooperate. This
can be done (unsuccessfully) through force, (successfully) by
creating/finding a common enemy, or (often successfully) by
a social structure that allows people to pursue their own
goals while contributing to "the common good".
A small example: people who want to exert power over
others should be capable of doing so, but in controlled
conditions (the typical, corrupt european politician is
vastly to be preferred over the absolute monarch).
Anybody basing their concept of society on the false
assumption that humans are by nature cooperative is
living in cloud-cuckoo land.
--
Stefaan
--
PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)
___________________________________________________________________
Perfection is reached, not when there is no longer anything to add,
but when there is no longer anything to take away. -- Saint-Exup�ry
------------------------------
From: [EMAIL PROTECTED] (Andrei A. Dergatchev)
Subject: Re: Slackware problem...
Date: Sun, 18 Jul 1999 16:06:28 GMT
Hi,
Glad to hear you've got Linux working at least on
part-time basis ;-)
>>Just curious - what's your HDD/motherboard ?
>Er... not sure :-) What's the easiest way to find out?
How about openning the chassis ? HDD is usually marked
somehow I believe, and mobos usually write something about
themselves during boot ?
>
>>As for a choice of bootdisk - what about no_pci.i,
>>just to try ?
>I've just tried it with no_pci.i, and finally installed Linux - hooray!
>I then restarted the computer with the boot disk it'd made, saw the
>login prompt, and - nothing! It had stopped again! NOOOO!!! AAARGH!!
> I restarted, managed to log in, did a few things, and then it stopped
>again! I think I'll have to give up if I can't get it working by
>Christmas... :-)
In fact I've also got a mysterious PCI-based problem,
but my case is exactly opposite to yours - Slack 4.0
was not able to recognize properly something with
my HDD's DMA, while 2 years back it wasn't a problem
for 3.4 !?
As for your current problems - I don't think you'd be
able to avoid kernel recompilation with *correct* PCI-
related settings. And I'm myself would be happy to dig
deeper in this area because of my problems. Please
inform me if you'll find anything interesting about that -
currently I'm still running 3.4 with upgraded to 2.0.36
kernel.
>--
>Philip Taylor
>philip @ zaynar . demon . co . uk
>http://www.zaynar.demon.co.uk/atr - Programming robots!
Good luck,
Andrei
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: Red Hat is Crap!!
Date: Sun, 18 Jul 1999 16:13:11 GMT
In article <7mrsj1$j93$[EMAIL PROTECTED]>,
Keven R. Pittsinger wrote:
>In article <[EMAIL PROTECTED]>,
> John Thompson <[EMAIL PROTECTED]> writes:
>> blah wrote:
>>
>>> X-Newsreader: Forte Agent 1.5/32.451
>>
>>> Get Slackware or Debian, they are the only pure distributions!
>>>
>>> Anyone who has used both Red Hat and either of these distributions
>>> knows this!
>>
>> This was from somebody using Windows to post...
>>
>> My take is that you may as well try seveal distributions.
>> Go with whichever one you want. If you don't like RedHat,
>> fine, don't use it. You have a choice.
>>
>> And what's all this about "pure distributions?" I suspect
>> this is only a concern if you're using linux for religious
>> reasons.
>
>
>RMS must love this guy.
One problem with RH is one cannot rely things work on other
systems too. E.g. there is now a RH glibc-2.1.2 ... great,
but up to now there is no glibc-2.1.2, no release. In
addition there are RH specific patches to the system libraries.
Good in one way, Applixware / glibc-version runs on RH with
glibc-2.1.x but up to now *only* on RH or in other words, the
RH glibc-2.1.x != official glibc-2.1.x, thank you very much.
No matter where I buy my HP from, I do not expect two machines
to behave different if they have been bought at two different
shops. I'd expect the same with a Linux distribution and right
now I'd stay away from RH as far as possible as a development
platform.
Minor note, I haven't got a distribution at all and it will
stay like that as long as that mess is getting worse, not
better.
Ta',
Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
------------------------------
From: [EMAIL PROTECTED] (Ilkka Ollakka)
Subject: Re: Can someone recommend....
Date: Sun, 18 Jul 1999 13:33:29 +0300
Reply-To: [EMAIL PROTECTED]
Viestiss� Sat, 17 Jul 1999 11:17:36 +0100, Andrew Arbon
<[EMAIL PROTECTED]> kirjoitti:
>I had a (brief) look at that yesterday, but I didn't think that it was
>an offline reader. As I am not in the US with free internet access, I
>need to read my news whilst not connected. It also seemed to
>automatically subscribe me to every NG my ISP carries. I don't want to
>have to unsubscribe to 40000 odd NG's.. Is there a way to prevent it
>doing that, or did I just misinterpret what it was doing?
>
>Thanks,
>--
>Andrew Arbon
Try using leafnode+slrn,I'm using system like this in news reading:
when I'm opening my PPP-connection,leafnode fetches new messages and new
groups from my isp's news server,but It only fethes those groups that you
want..
sorry my poor english,but hope that you get the point..
Ilkka Ollakka
--
--
DeVries's Dilemma:
If you hit two keys on the typewriter, the one you don't want
hits the paper.
------------------------------
From: [EMAIL PROTECTED] (John Reiser)
Subject: Re: Solving the 1024 cylinder LILO problem
Date: Sun, 18 Jul 1999 16:52:21 GMT
[... snip ...]
>> All the popular Linux
>> distributions are broken in this way.
>
>correction - all the popular *BIOSes* are broken this way. this
>problem is not operating system specific. this is a boot problem
>caused by deficiencies in the PC architecture.
>
[ .. snip ...]
>*BIOS* has to load you operating system you know. and in
>order to load it, *BIOS* has to be able to access it. this is a
>*BIOS* problem. take this to american megatrends and the other *BIOS*
>manufacturers.
>
My experience is that American Megatrends is one of the good guys. I
have a SuperMicro P5STE board with an AMIbios that displays dates of
052797 and 071595, and the BIOS is designed to support disk access to
at least 2**31 sectors of 512 bytes (total 2**40 bytes, which is 1024
gigabytes). I have not tested it that far :-) See
ftp://ftp.teleport.com/pub/users/jreiser/mbr03.tgz and e2boot4c.tgz
for GPL booting software that can detect and use such BIOS support if
available. There is an argument that LILO itself is behind the times
for not including similar support.
See http://www.phoenix.com/pcuser/bios.html#Q5.7 :
"AwardBIOS's dated July 1, 1994 and later support hard disks up to
8.4GB in size. Some AwardBIOS implementations after May 1, 1997
support hard drives larger than 8.4 GB - it depends on whether the
[motherboard] manufacturer chose to implement this feature."
See http://www.phoenix.com/products/specs.html , and particularly:
http://www.phoenix.com/products/specs-edd11.pdf
BIOS Enhanced Disk Drive Specification, Version 1.1, May 9, 1995.
How the BIOS deals with more than 1024 cylinders.
http://www.phoenix.com/products/specs-edd30.pdf
Latest version, with extensions for managing drive letters when
adding new drives, etc.
http://www.phoenix.com/products/specs-bbs101.pdf
BIOS Boot Specification, Version 1.01, January 11, 1996.
"Plug and Play" BIOS uses 8 bytes before the partition table.
Register dl passes the drive number for booting.
So [at least some of] the BIOS developers have been at least
approximately abreast of the increase in capacity. If there is a
restriction in hardware purchased in 1995 or after, then it is due to
an _economic_ choice of the motherboard manufacturer and consumer.
The BIOS to support big disks was available, and a discriminating
consumer could have bought a PC that included it.
Somewhere around 8GB or less is the practical limit of 3.5-inch wide
by 1-inch high ("half height") hard drives which use only thin-film
magnetic heads ("Winchester" technology). Most of the dramatic
increase in capacity in recent years has been due to the use of
magneto-resistive ("MR") recording technology, which enables practical
storage of several times as many bits per square inch of active
platter surface. There were researchers and visionaries who
knew/predicted this before the early 1990's. But they were few, and
the economics strongly favored riding the magnetic-only horse, which
has a history _longer_ than that of the integrated circuit.
[EMAIL PROTECTED]
------------------------------
From: Stephen Chadfield <[EMAIL PROTECTED]>
Subject: Re: Wordperfect and Linuxmall Confusion!
Date: Sun, 18 Jul 1999 13:53:05 +0100
ryan wrote:
>
> I recently ordered a so called 'download version' CD-ROM of Corel
> Wordperfect 8 from linuxmall.com.
>
> Linuxmall claims that this is the free downloadable version of
> wordperfect. However the CD-ROM is fully packaged in Corel's jewel case
> and printed CD for the full personal edition. After trying to install
> WordPerfect, I have come to believe that this is the actual Personal
> Edition, even though I only paid $2.99.
The free version you can download is called the "Personal Edition". See:
http://linux.corel.com/linux8/index.htm
--
Stephen Chadfield
http://www.aquamarine.demon.co.uk/
------------------------------
** 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
******************************
