Linux-Advocacy Digest #704, Volume #29 Tue, 17 Oct 00 09:13:05 EDT
Contents:
Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum (Chris Sherlock)
Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum (Chris Sherlock)
Re: Because programmers hate users (Re: Why are Linux UIs so crappy?) (Roberto
Alsina)
Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum (Chris Sherlock)
Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum (Chris Sherlock)
Re: Is there a MS Word (or substitute) for Linux? (Brian Langenberger)
Re: Is there a MS Word (or substitute) for Linux? (Jem Berkes)
Re: Is there a MS Word (or substitute) for Linux? ("MH")
----------------------------------------------------------------------------
Date: Tue, 17 Oct 2000 22:52:40 +1000
From: Chris Sherlock <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum
Chad Myers wrote:
>
> "Nick Condon" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Chad Myers wrote:
> >
> > > You're comparing a relatively new (although based on archaic technology)
> >
> > This one always irritates me, the Windows people say it about Linux, and the
> Linux
> > people say it about Windows.
>
> Well, the Linux people seem to not understand the difference between
> Win9x and WinNT/2K (which only further proves my theme that Linvocates seem to
> have not much clue about the state of the rest of the computing industry).
>
> WinNT/2K is not based on DOS. Period. DOS (cmd.exe) in NT is merely a tiny
> subsystem for compatibility, just like POSIX subsystem and the OS/2 subsystem.
This is very true. The underlying architecture of NT/2000 is an
interesting one. Having a HAL is actually quite a good idea, however how
does this effect performance of the O/S?
> Linux, however, is almost entirely based on it's Unix predecessors (which isn't
> necessarily a bad thing). Linux is somewhat reminiscent of Win9x (yes, I know
> Linux has a better kernel, etc that's not my point). Win9x is hack upon a
> hack upon a hack upon an old 16-bit non-multitasking OS. By way of comparison,
> Linux is a hack upon a hack upon a hack upon a 30 year old archaic OS that has
> none of the features of a modern OS.
What features would they be? Multi-tasking, multi-user, networkable, SMP
support? These are *all* age old O/S concepts.
Considering that Linux's kernel was built from scratch, but has
successfully used Unix filesystem and proven Unix techniques I am
intrigued by what modern features you are referring to.
> Linux should just be called the Unix
> modernization project. Except all it does is add more instability and more
> inconsistency.
I have found Linux to be extremely stable, and in most things extremely
consistent. Again, may I ask what inconsistencies you are referring to?
> > But unless you're talking about the wheel or fire, all technology is based on
> older
> > technology.
>
> But how deeply is it rooted? That's the question.
I don't feel that Linux technology is rooted in archaic technology. It
runs (very successfully) web servers, DNS servers, news servers and mail
servers. You scale it using Beowulf and/or an SMP kernel, if you so
desire.
I find Unix concepts to be extremely elegant.
Chris
> -Chad
------------------------------
Date: Tue, 17 Oct 2000 22:59:30 +1000
From: Chris Sherlock <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum
"." wrote:
>
> In comp.os.linux.advocacy Chad Myers <[EMAIL PROTECTED]> wrote:
[snip]
> > and you can
> > run most OS/2 2.x applications with the OS/2 subsystem (which is what it was
> > designed for).
>
> No, you cant. You can however hack and slash OS/2 programs until theyll talk
> to the OS/2 layer a little, but god help you if they want direct access to
> hardware. (as most useful OS/2 software does; see call managers, video channel
> managers, etc.)
I agree with a lot of what you have to say, er, dot (?) however isn't
the whole *point* of NT's hardware abstraction layer to *stop* the
direct manipulation of hardware? I always assumed that that is what it
was there for. Of course, you don't access hardware directly through
Linux apps either - you write a device driver with well-defined
interfaces, however I just wanted to make this point.
Of course, I do realise that you were talking about OS/2 here. Possibly
OS/2 does things differently than what I know of. In which case I will
freely admit I may be incorrect in something I said here.
Chris
> > IIRC the POSIX version in NT 4 and also Win2K is something like 1.2. There
> > are updated POSIX subsystems for NT/2K from 3rd parties and even from
> > Microsoft in their Services For Unix.
>
> Yes, and their services for unix arent bad, not bad at all.
>
> They however would work fine without NT's POSIX layer---highlighting its
> uselesness except in the certification department.
>
> -----.
------------------------------
From: Roberto Alsina <[EMAIL PROTECTED]>
Subject: Re: Because programmers hate users (Re: Why are Linux UIs so crappy?)
Date: Tue, 17 Oct 2000 09:31:34 -0300
El lun, 16 oct 2000, Richard escribió:
>Roberto Alsina wrote:
>> >Having objects that are "classes" of other objects is ridiculous. Having
>> >them be magical entities that aren't addressable as objects (yet still
>> >exist somehow someway) is insane.
>>
>> So, in short: having a class is not a necessary characteristic of an object.
>> So, not belonging to classes doesn't prove classes are not objects.
>
>What the fuck are you talking about? Do you realize that what you
>just wrote is a complete non sequitur?
You seem to be unaware of an artifact called explanation.
>> So, your previous "what class are they supposed to be?" was logically
>> ineffectual, since a language could have objects that belong to classes and
>> objects that don't.
>
>No, it can't bozo. EVERY object belongs to some conceptual class. Classes
>might not exist as language constructs but that doesn't stop them from
>existing in people's minds.
I naïvely assumed you were using a constant frame of reference. In that case,
classes belong to the class "class", in my mind.
> And if a language chooses to have classes as
>language constructs then these MUST match the conceptual classes that
>people build of the system.
Says who? A language needs not be so coherent.
> The C++ decision to have some classes be part
>of the language and others not is ludicrous.
Yet, C++ exists. Therefore, it is not correct that languages MUST, is it?
>> So much for your vaunted logic skills.
>
>Logic is beyond you.
Indeed it is. Logic goes much further than I have ventured. Apparently,
however, you believe logic is not beyond you. Which is appalling.
>> And before you say I am stupid: otice how I don't claim that classes are
>> objects, I just point out the inadequacy of your own attempt at doing that.
>
>You are stupid, there's no disputing that. The only question is the
>precise depth of your stupidity.
I will indeed not dispute whether I am stupid or not. I am not impartial.
>There are bound to be inconsistencies in my stated position in cases,
>such as this one, where my position has evolved over time. And yet
>you're too stupid to find any of these inconsistencies.
"Over time" seems to mean "a day". If your position changes, just say, "oh,
what you quoted, I was wrong there".
>If a language has a class construct at all (and there is no reason
>why it should) then people should be able to access and manipulate
>this construct as easily as any other construct in the language.
Languages need not be introspective.
> It's as simple as that. If classes exist in the language then they must
>be objects, just as much as if classes exist in the language then
>everything must belong to some class.
You assume the language tries to be completely object oriented. Such is not the
case with, for example, C++.
--
Roberto Alsina
------------------------------
Date: Tue, 17 Oct 2000 23:31:24 +1000
From: Chris Sherlock <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum
mlw wrote:
>
> Mike Byrns wrote:
> >
> > 2:1 wrote:
> >
> > > > Actually, it's not. Win2K is way more stable than Linux. FUDsters like
> > > > to bash NT for stability and then claim that Linux is stable, which in
> > > > and of itself is a joke.
> > >
> > > Chad, you're talking horseshit again. Win2K hasn't been around long
> > > enough to proove if it is more stable than Linux. The longest Linux
> > > uptimes are longer than the total life of Win2K.
> >
> > Where are those numbers? We're not talking peak are we :-) I've had an NT
> > 3.51 box running for 3 years. We decommissioned it.
>
> By running, do you mean 24x7 with no reboots?
>
> When UNIX (Sun, FreeBSD, Linux, HP, etc.) people say "running" they
> mean, on and up. It has been my experience when NT people say running,
> them mean with a particular install and periodic service packages and
> reboots.
This is my experience also. At my last firm, they a) could not afford to
run Windows NT because of the immense cost of licenses, and b) it truly
was too unstable and unreliable. These machines needed to run a SQL
server and DHCP servers. They also had to be the intermediary to the
outside world (admittedly through a router, but they had to perform NAT
and IP filtering before it reached the router). It also had to be able
to do basic snmp trapping, and it had to run a web proxy but that's
another story.
Could you imagine trying to do this on a Windows NT box? Could you
imagine trying to remotely monitor AND maintain such computers?
Unfortunately the firm had buggy software which would die or degrade
very badly. On the Linux box you would simply ssh onto it, kill and
restart the process. Remote upgrades? Sure, you needed to have someone
on-site in case something weird happened, but then again, so what? On a
Windows NT computer it would be far too expensive to implement, far too
unstable and far too slow for the hardware that it was running on. If a
process did get out of hand, how would you kill it remotely? We are
talking international monitoring here, btw. How would you restart it?
Get someone to go to the computer and reboot the computer? This is not a
very viable option, I am afraid.
In short, Windows NT was not even an option.
> I could be wrong, but it seems to me an uptime of three years requires a
> very good infrastructure with no power glitches. Also, Windows NT 3.51
> had AFAIR some pretty notable memory leaks which could cause problems if
> not rebooted.
>
> Seriously, Windows NT (at its core) is a very good OS, it is based on
> VMS after all. It is the Win32 subsystem and the kernel code required
> there in, which causes instability. Win32 was not designed to be a
> secure robust environment, so as a requirement of implementation of it
> they (MS) had to make compromises in the NT kernel space which makes it
> unsuitable for many applications.
>
> The first indication that Microsoft was not serious about the enterprise
> was Windows NT 4.0, when they moved GUI code down into kernel space.
> Windows 2000 took this trend to new disgusting levels.
Crap. I was going to point out that you were wrong, but hey, you know
what? MS really *did* put the GUI back into the kernel area. They even
say so themselves at
http://www.microsoft.com/technet/winnt/winntas/technote/ntunixvw.asp I
quote: "(Editor's note: With Windows NT version 4.0, the GUI system was
put "back" into the kernel for display performance considerations.)"
That *really* sucks! I admit that I am no Windows NT fan, but no
*wonder* people don't use it for mission critical things without needing
to do advanced load-sharing methods. I can't believe that a GUI would
play such an important part of a server O/S!
> I am concerned that Linux is making the same sorts of compromises,
> however, I am confident that it will always be "optional."
Linux has a DRI, however this actually turns out to be quite useful, and
AFAIK I beleive that this is optional when building the kernel (someone
correct me if I'm wrong!) The whole GUI system is *not* going into a
kernel subsystem.
> --
> http://www.mohawksoft.com
BTW, while I was doing some research for this post, I came upon the
following pearl from Microsoft:
"few UNIX kernels today are dynamically linked. So relinking and
reloading the kernel is often required, which is another systems
management task."
(http://www.microsoft.com/technet/winnt/winntas/technote/ntunixvw.asp)
Oh really? This is a little bit rich coming from Microsoft! Frequently,
when I installed a new piece of hardware under Windows NT (I tried to
install two network cards, a 3Com and another generic one which I can't
remember what brand it was. Apologies for this) I had to restart the
computer *several* times.
Incidently, I understood that the process of loading/unloading kernel
modules in Linux was dynamically linking. Surely other Unixes do this?
Chris
------------------------------
Date: Tue, 17 Oct 2000 23:34:58 +1000
From: Chris Sherlock <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 mired in delays as Compaq warns of lack of momentum
Which network group are you talking about?
Chris
[EMAIL PROTECTED] wrote:
>
> On Wed, 11 Oct 2000 10:16:25 -0500, Spicerun <[EMAIL PROTECTED]>
> wrote:
>
> >Actually, I'm not going to spend the large amount of money just for an upgrade of an
> >OS that will probably not work with my ,webcam or gaming card' (Actually I don't own
> >a webcam or gaming card), or network cards, or jukebox CD or even regular CD. In
> >fact, if WinME full version install was any indication, I think I'll stay away from
> >future MS upgrades.
>
> I wouldn't either.
> I'm running Win98SE for that very reason. For what I do, Win2k is a
> poor solution. Nowhere near as poor as Linux, but a poor solution just
> the same.
>
> >I will see Win2K soon enough at work, but I do notice with some chagrin that our IT
> >department is still trying it out, and are kinda 'pissed off' that they have to
> >evaluate it for a longer period of time to work out the issues because it won't work
> >out of the box with our particular network. I know of at least 3 IT administrators
> >at work that wish Win2K would go away.
>
> Can't comment here, except that Win2k supported my P2P fine along with
> ICS as well. So did Winme and WinSe.
> Linux supported the Network as well, even easier than Windows in fact
> (no screwing with adding protocols and bindings and such along with
> reboots, it just worked), but I never did figure out how to set up ICS
> under Linux, nor a decent firewall that has a rules assistant etc like
> Norton.
>
> Seems to be a common question in the network group.
>
> claire
>
> >====================
> >Standard Disclaimer: My opinions are my own and do not represent the views or
> >opinions of my employer.
> >
------------------------------
From: Brian Langenberger <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Is there a MS Word (or substitute) for Linux?
Date: 17 Oct 2000 12:43:49 GMT
In comp.os.linux.advocacy Harry Lewis <[EMAIL PROTECTED]> wrote:
<snip>
: If all you have is a hammer, everything looks like a nail.
: If all you have is a typesetting program, everything looks like a
: typesetting problem.
And if all you have is a word processor, everything looks like a
page layout problem.
: Problem is, everything isn't a typesetting problem. Typesetting is the
: business of printers. Word processing is what users do.
Typesetting is TeX's problem. Picking the right fonts and margins
and page numbers is LaTeX's problem. Generating content is the user's
problem.
Word processors are nice for quick and ugly disposable documents,
but don't expect me to write anything of value with one.
------------------------------
From: Jem Berkes <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Is there a MS Word (or substitute) for Linux?
Date: Tue, 17 Oct 2000 07:45:33 -0500
Powerful wordprocessor? Word Perfect for Linux (it's free)
http://linux.corel.com
------------------------------
From: "MH" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Is there a MS Word (or substitute) for Linux?
Date: Tue, 17 Oct 2000 09:07:25 -0400
"mlw" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> You can download Star Office. I used to use Applix, but I got their last
> version and it wasn't as good as their previous version. SO 5.2, OTOH is
> as good as MS office in that it doesn't crash like MS office, and it
> does not take down the OS like MS office.
>
> It has no problems reading MS office files, at least as far as I've
> seen.
I've been using office since 4.3. I've NEVER had it "take down the os".
NEVER.
Star office == to ms office? I'm not attacking or being personal, but that's
the most ridiculous statement coming from someone who works in the
"industry" I've ever heard. There is NO comparison between the two products.
Either on stability, feature set, or programmability.
If these products were equal, (for the sake of argument) would Sun have to
give it away?
No. They give it away for the same reason most open source programs are
given away.
They can't compete with the retail equivalents. Simple marketing. MS can run
ads until they empty their bank. Sun could do the same and STILL NOT CHARGE
for SO. SO would not survive in that market because it's an inferior
product. The standard has been set, the people educated in its use, and we
have entrenchment... "saying" one is better isn't going to get it. End of
story.
You have no problem reading MS office files? There must not be *much in*
those office files, huh?
Excel provides 100's of functions, and the entire via language, which many
experienced spreadsheet designers use. SO supports this?
Word's tables, revision features, html formatting, comments, and many other
things...SO supports these?
Access with either DAO or ADO, forget it....
PowerPoint? again, forget it.
Yes, it can handle very simple conversion chores from MS office format, and
That's IT.
Saying that you've never had a problem tells me you don't import much from
one to the other.
And again, this is linux advocacy. You many have much knowledge in your
field of expertise, but you don't use computers as real computer users do.
They have to have compatibility with one another. Not a tower of babble.
Like it or not, the standard is Office. Spreading lies and FUD does not help
Linux's cause. If you had said simply that "SO can open many standard office
formats, but you may have trouble with some of them-- the conversion is not
perfect", would have been different. But no, more Linux BS.
Plus, I don't understand the following statement, "SO is as good as MS
office in that it doesn't crash like MS Office". What the hell does this
mean? SO doesn't crash? That's pure BS. SO crashes less then MS office?
That's completely unsupported by any facts you bring to the table. (none).
OK, let us say the SO does crash less than MS office,would that and that
alone make it better than MS office? But WE CAN'T say that, can we? Because
we offer no evidence that that is the case. This is what I DETEST about
linux advocacy. If I have a complaint about Linux, or a linux application. I
say why, in no uncertain terms what it is I dislike, and will be happy to
give particulars to back it up. You, and your cola partners in crime, on the
other hand, make assertions like the above. You have nothing to back it up
but the weight of your signature and the ubiquity of your presence on
usenet. That and 50 cents might buy you a paper. Meanwhile, MILLIONS of
people are using MS office every minute of every day. And yes, they DO get
work done! That's the reality of what is facing Linux. That is THE desktop.
You can dance around it, but if your argument is going to always resemble
what you proffer above, you're in deep shit people.
------------------------------
** 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
******************************
