Linux-Advocacy Digest #698, Volume #27 Sat, 15 Jul 00 14:13:06 EDT
Contents:
Re: one step forward, two steps back.. ("MH")
Re: Linsux as a desktop platform (Gary Hallock)
Re: Some Windows weirdnesses... ("James")
Re: This thread has needed a new name from the beginning (SomeOne Else)
Re: Linsux as a desktop platform (tinman)
Re: one step forward, two steps back.. (Cihl)
Re: Linsux as a desktop platform ([EMAIL PROTECTED])
Re: Linsux as a desktop platform ("Christopher Smith")
----------------------------------------------------------------------------
From: "MH" <[EMAIL PROTECTED]>
Subject: Re: one step forward, two steps back..
Date: Sat, 15 Jul 2000 13:11:43 -0400
"Cihl" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
Well, this is obviously a troll, but i'll answer anyway.
Actually I wasn't trying to troll. I think Linux is fantastic. I just don't
like some of the directions the major distributors are taking. I realize I
can take that or leave it.
Thank goodness for that!
Oh, and thanks for the very reasonable counterpoint devoid of flame.
'--------------
I don't think the installation programs are emulating Windows in
particular. What the distributions want, in general, is to provide for
easier installation procedures for new users. The older installation
procedures were good in itself, but they tended to scare off these new
users.
'--------
I agree now that I think about it. They aren't emulating it at all. If they
were, the installation would work a little better. :-)
'-------
When you say that you're accustomed to older versions of Linux, even
mentioning these problems seems strange at best.
For instance, i don't recall Linux ever mentioning anything on boot
about modems. They either work or they don't. And how can the sound
work sometimes and sometimes not? In Linux it either works always, or
never. Three years ago installed programs were never mentioned on any
menus.
'--------
Red hat has included 'value added' features. One of these is checking for
new hardware during boot up. A red hat screen pops up (color) that informs
me that the device (serial modem) is no longer present and will be removed
from the system. But, I have the choice to 'ignore' this and the device
isn't removed. The serial modem is where it is every boot. Jacked into the
com2 serial port.
The sound I can't explain. Doesn't really bother me. I don't use it other
than to play audio CD's...which I can live without. Just seems funny to me,
I get no error messages when it doesn't function. At least none that are
obvious. But then I'm not that anal about reading all the system messages
and logs. I'll leave that tedium to those with nothing better to do with
their workstation.
As for menus. I remember installing RH 5.1 and with only 2 exceptions,
everything I installed in the way of programs showed up on any window
manager's menu. Now it seems not. And the menus seem very disorganized.
'----------------
Drag 'n drop support has been completely rewritten just recently, and
both Gnome and KDE are going to use it in their next incarnations.
(I'm talking about XDnD, of course)
'---
Perhaps I'll try it when it's done.
'--
If you run the same WM's you did three years ago, you will have the
same amount of memory usage, too. NN hasn't changed much, either. It
added some support for things, and some instability with it. Looks
like somebody is exaggerating here. :)
'------
I would, I just don't have time to figure out how to 'de-gnome' my
interface.
When I switch to say, windowmaker that damn Gnome taskbar is still there
with WM laid over top of it. The memory? I'm not making it up. Unless free
is lying, I'm swapping when I never used to. Of course I think RH has
installed every daemon known to nix, of which 50% I don't need. I haven't
gotten around to trimming the modules.
'--------
But you can still have it the way you had it three years ago. IceWM
still exists. Linux is a *fully modular* system. You can install it
anyway you want it. It doesn't require any more work than it did three
years ago.
'----
I agree. Just wondering if I can *kill* Gnome. It seems entangled into the
very knooks and crannies of any window manager I choose to run (=
take care.
------------------------------
Date: Sat, 15 Jul 2000 13:21:38 -0400
From: Gary Hallock <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
The Ghost In The Machine wrote:
> I prefer PMT, but the way microprocessors are going (how many regs
> does a Pentium III have?), it's going to get a high overhead
> to save everything on a context switch. [+]
>
There are solutions to that problem. I'm not familiar with the design of the
Pentium III, but I have worked on processor designs that have multiple sets of
registers in the hardware. Architected registers then become virtual registers and
a context switch simply becomes a matter of switching register sets. Of course,
at some time, when there are no free register sets, you must save to storage.
But it can reduce the average context switch time.
> >
> >>But, as has been explained to you many times, that is simply not
> >>true. PMT, by its very nature, can respond faster than CMT. This is
> >>because, when the user clicks on the mouse, a PMT system can grab control
> >>away from the currently running process and give it to the user interface
> >>process.
>
> Not quite; PMT can do this, but it would require quite a bit
> of engineering and a (mostly) idle system. Assume that the mouse is
> serial, which means serial interrupts every time it moves. (Presumably
> a bus mouse is similar.) When the mouse moves, the kernel is interrupted,
> causing a context save; the kernel then looks around for any process
> waiting on the mouse. If there is one, it puts it on the ready queue
> on a PMT system (and possibly on a CMT system, as well). The kernel
> also reads the characters from the mouse at some point, as well
> (it doesn't bother to translate them into coordinates, however;
> that's someone else's job).
>
> If there's nothing else running, the first process is fetched from
> the ready queue, a context *switch* is done [*], and the new process
> is now running. If there is something else running, the kernel
> gets to play "who gets the CPU", and there are various schedulers
> available, of which I don't know a lot about beyond the fact that
> the Amiga didn't have dynamic reprioritization and a lot of other
> systems do (i.e., the longer a process waits, the higher its
> dynamic priority gets, and the more likely it will get at least a
> small piece of the system). There are also issues with swapping; if
> the process wanting the mouse is swapped out (because other processes
> needed the RAM), and the mouse is moved, things will get slightly
> slow as the kernel shuffles pages around.
>
> Also, the kernel can forcibly context switch on a busy system,
> every quantum (10 ms? 100 ms? I'm not sure; it may depend on
> the system), shoving the running process back onto the tail end
> of the ready queue and picking up another process therefrom.
>
> So -- just because the mouse moves, doesn't mean the right part of
> the system will notice. It depends on a lot of factors.
> Note that on Linux systems, the X server reads /dev/mouse (a symlink
> to a character special file which throws out raw bytes -- could
> be ttySn, could be psaux) and interprets them as mouse actions;
> therefore, if X is paged out, bye bye performance.
> Unix systems are similar, although the names will be quite
> different depending on HP-UX, Solaris, AIX, OSF/1, Ultrix,
> Tru64, SCO Unix, and even Xenix.
>
> (X being paged out wouldn't be that often, though, because
> everybody has to go through X to make marks on the screen.)
>
> However, at least the kernel has the option of forcibly moving
> things around; on a pure CMT system, this isn't really possible.
>
All very true. Which is why a real time OS needs to do some things differently,
including, for example, pinning memory for a high priority process. But for user
interfaces, the time to process the interrupt and handle the context switch on a
PMT OS is very small in human terms. CMT makes it likely that the time will
actually be noticeable by a human being.
Gary
------------------------------
From: "James" <[EMAIL PROTECTED]>
Subject: Re: Some Windows weirdnesses...
Date: Sat, 15 Jul 2000 19:27:49 +0200
Virgo,
Why are you bothering to discuss a mickey-mouse OS like win95 in this NG.
If you need stability use Win2k, or Linux (if you don't need serious desktop
apps).
James
"V'rgo Vardja" <[EMAIL PROTECTED]> wrote in message
news:8kmlmh$1bh2$[EMAIL PROTECTED]...
> Hi,
>
> I've been lurking in c.o.l.a for a while, observing the constant battle
> between Linvocates and Windows lovers. The main point of the Winvocates
> seems to be 'Linux doesn't support my hardware or my Word documents and
> GIMP sucks compared to Photoshop because GIMP is called GIMP'.
> Linvocates retaliate with 'Windows crashes and needs a reboot every 42
> seconds and every minor version changes driver format'. For myself, I
> like Linux. And I tolerate Windows. I won't go throwing it out of the
> nearest airlock, but I sure as hell don't think Windows is as perfect as
> certain you-know-whos try to convince us.
>
> See, Winvocates, you've been trying to get the Linux community to
> acknowledge the imperfections of Linux. Now I'd like to hear what *you*
> can say about the couple of weird Windows quirks.
>
> My home computer is a 486DX/33 (16M memory), which dual-boots Win95 and
> Slackware 7.0. Linux runs perfectly; Windows has a weird bug: it refuses
> to cold-boot correctly. If I turn the power on and let Windows boot by
> itself then the boot process hangs before the GUI shows up. The only thing
> I can do to un-wedge it is another cold boot. So now if I want to use
> Windows I'll first have to hit F8, boot to DOS mode, run Norton Commander
> and then hit *CTRL* *ALT* *DEL* and after that Windows warm-boots
> correctly. I've tried two versions of Win95 and several reinstalls, and
> it still has that weird problem.
>
> I work in a private school as a general all-around friendly
> neighbourhood computer guy, BOFH and/or webmaster. Which means I also
> take care of our computer labs (Windows 95 networked in a WinNT domain).
> So here's a couple of, er, interesting Win95 *features* I've encountered
> so far.
>
> Have you ever seen Win95 forget the hardware it sits on? Well, I have.
> Sometimes a teacher will come and tell me that a soundcard isn't
> working, or a computer is unable to log on to the network. So I'll have
> a look and find that the soundcard or network card definition is missing
> or broken in the System Properties. And yes, we have denied ordinary
> users access to most Control Panel applets with system policies, so it
> could not have been the fault of children removing the devices.
> Reinstalling drivers works with the soundcards but to get the network
> adapters going again I have to remove the software with Add/Remove
> Programs applet and *then* reinstall. And we're *not* talking broken
> hardware.
>
> When you log on to Microsoft Networked computer you are asked whether or
> not you want to save your personal preferences. If you say 'yes', your
> user.dat gets saved to c:\windows\profiles\your_login_name\user.dat. If
> you say 'no' then your profile will *not* be saved - instead the changes
> you might make in the interface configuration will be saved as the
> default configuration, for instance your desktop background and screen
> saver properties. So people would set up a password-protected screen
> saver which happily kicks in after they log off. No one else could use
> the computer unless we did a physical reset. *CTRL* *ALT* *DEL*
> combination doesn't work and there is no administrative password (unlike
> in WinNT where administrator's username and password can unlock a
> password-protected session) and it is not possible to log on to the
> computer over the network like for instance with Linux. After a while I
> finally found a program that can access user.dat on a remote machine and
> crack the password (thank goodness for Microsoft's lousy screensaver
> password encryption!).
>
> Win95 does not do a good job at protecting its underlying hardware. I've
> seen countless of occasions where a game fscks up monitor refresh and/or
> screen resolution so that the screen flickers like something from the
> late 80's. Ever seen a 320x200 screen resolution with 800x600 virtual
> desktop? Not good.
>
> Our system policy makes it unable for our users to access the computers
> without logging on to the network (no pressing *ESC* and gaining access
> to the computer, that is). So the younger children who do not have their
> own accounts yet can only use the computers if a teacher logs them on.
> Unfortunately the kids found a way to bypass that restriction by
> replacing the domain name in login dialog with a bogus one - Win95 tries
> to log on to that non-existent domain, gets no answer but nevertheless
> assumes that it's OK to allow the login. Officially there was nothing I
> could do to stop them; there were no patches we could find from Microsoft
> and no way to get rid of the Domain: field of the login dialog. So I did
> something that probably goes against Microsoft's licence - I used the
> resource editor from Visual C++ to disable (gray out) the Domain: field
> within a Windows DLL... This got rid of that problem. (talk about
> configurability...)
>
> Then there is the issue of filesystem quotas - something Win95/98/NT
> does not have. Our network admin has set the disk space limit on the
> home directories to 5 megabytes for ordinary users... unfortunately the
> NT server does not yet have a way of enforcing the quotas so he checks
> the sizes of the users' home directories every once in a while. So far
> the record is five bloody hundred megabytes of mp3's in one person's
> home directory. And the admin does a nightly backup from the partition
> on the server where home directories are kept... Now someone tell me
> that Windows rules compared to Linux. Hah!
>
> And I won't even mention the number of Windows crashes. My own
> workstation (which is the only one right now that runs Linux) has
> probably the lowest amount of crashes in the whole school complex, and
> those crashes only happen because it also double-boots Windows. The
> only times I've managed to crash Linux was when I tried to compile my
> very first kernel (oops! modularised IDE disk support and ELF binary
> support... guess what that meant) and when I tried to get a scanner to
> work with the poor excuse it had for a SCSI card.
>
> So, dear winvocates, you can now try to present your solutions to the
> problems I just described, which would not exist if the computers were
> running Linux. And please, don't bother with statements like 'LINUX
> SUXXX' and 'Linux doesn't support my printer' or 'Buy Win2k'.
>
> Happy hacking,
>
> Virgo
>
> --
> ERROR: CPU not detected. Emulating.
> -Win2k
------------------------------
From: [EMAIL PROTECTED] (SomeOne Else)
Crossposted-To: alt.sad-people.microsoft.lovers,alt.destroy.microsoft
Subject: Re: This thread has needed a new name from the beginning
Date: Sat, 15 Jul 2000 17:29:24 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 15 Jul 2000 02:28:10 -0400, T. Max Devlin <[EMAIL PROTECTED]>
wrote:
>Said Bob Lyday in alt.destroy.microsoft;
>>"Aaron R. Kulkis" wrote:
>>>
>>> "T. Max Devlin" wrote:
>>> >>
>>> >But I've noticed that to
>>> > most people these days "proper administration" and "competent end-user"
>>
>>Some people consider me to be a computer geek. I get about 1-2 blue
>>screens a day and have to reboot Windows on a daily basis. I am
>>incompetent?
>
>Not as far as I know. You're a computer geek. You just said so. Power
>users typically get *more* problems with windows then timid normal ones.
>
Power users tend to have more problems then regular ones because they
actually use the software. A person that drives the latest Ferrari but
never faster then 15mph is not going to notice any where near as many
flaws as a person who constantly pushes 65mph.
So why are these "15mph" users suffering from all the disadvantages
of large bulky software to do their trivial instead of smaller leaner
software more suited to their needs.
------------------------------
From: [EMAIL PROTECTED] (tinman)
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sat, 15 Jul 2000 13:30:14 -0400
In article
<[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> On Fri, 14 Jul 2000 15:44:40 GMT, Passenger Pigeon
> <[EMAIL PROTECTED]> wrote:
>
> >In article
> ><[EMAIL PROTECTED]>,
> >[EMAIL PROTECTED] wrote:
> >
> >> On Fri, 14 Jul 2000 15:11:25 GMT, Passenger Pigeon
> >> <[EMAIL PROTECTED]> wrote:
> >>
> >> >In article
> >> ><[EMAIL PROTECTED]>,
> >> >[EMAIL PROTECTED] wrote:
> >> >
> >> >> Poor Mac. When did the Mac get 256k or 512k?
> >> >
> >> >Now, I know this is a difficult logical leap, but somehow I just assumed
> >> >that the Mac 512k had, oh, somewhere in the ballpark of 512k of memory.
> >>
> >> No shit. But when?
> >
> >http://search.metacrawler.com/crawler?general=mac+models&method=0&redirec
> >t=web&rpp=20&hpe=10®ion=0&timeout=0&sort=0&format=beta99&theme=classic
> >&refer=mc-search
>
> When? Can you give me a year? That link leads to many other links,
> having nothing to do with the subject at hand. I'm sure you know -
> just tell us.
The Mac128 came out 1/24/1984.
The Mac512 came out 9/10/1984.
IIRC the Mac512 shipped as soon as the 256 mem chips were available in quanity.
--
______
tinman
------------------------------
From: Cihl <[EMAIL PROTECTED]>
Subject: Re: one step forward, two steps back..
Date: Sat, 15 Jul 2000 17:39:54 GMT
MH wrote:
> Actually I wasn't trying to troll. I think Linux is fantastic. I just don't
> like some of the directions the major distributors are taking. I realize I
> can take that or leave it.
> Thank goodness for that!
>
> Oh, and thanks for the very reasonable counterpoint devoid of flame.
> I agree now that I think about it. They aren't emulating it at all. If they
> were, the installation would work a little better. :-)
Well, to be quite honest. Installation procedures aren't very good in
Linux. And the same goes for Windows. The only good example of a
capable installation procedure is BeOS. (Too bad BeOS doesn't do much
else, though.)
> Red hat has included 'value added' features. One of these is checking for
> new hardware during boot up. A red hat screen pops up (color) that informs
> me that the device (serial modem) is no longer present and will be removed
> from the system. But, I have the choice to 'ignore' this and the device
> isn't removed. The serial modem is where it is every boot. Jacked into the
> com2 serial port.
I believe you can turn the hardware detection bit off on boot. I can't
remember how in RedHat, though. In Mandrake you just run DrakConf.
> The sound I can't explain. Doesn't really bother me. I don't use it other
> than to play audio CD's...which I can live without. Just seems funny to me,
> I get no error messages when it doesn't function. At least none that are
> obvious. But then I'm not that anal about reading all the system messages
> and logs. I'll leave that tedium to those with nothing better to do with
> their workstation.
Well, i can certainly tell you that's not supposed to work that way.
Maybe it has something to do with a BIOS-setting. PnP-OS?
> As for menus. I remember installing RH 5.1 and with only 2 exceptions,
> everything I installed in the way of programs showed up on any window
> manager's menu. Now it seems not. And the menus seem very disorganized.
Personally i avoid RedHat like the plague. My best experience with
Linux installation and configuration lies with SuSE. Maybe you should
try that one as well.
[referring to dnd]
> Perhaps I'll try it when it's done.
KDE2 will certainly use this new technology. It's be out in a maximum
of 8 weeks from now. The first distributions carrying KDE2 will come
out next fall or so. For some wow's, look at the KDE2-screenshots from
http://www.kde.org.
I don't know about Gnome. Can anybody else fill this in?
> I would, I just don't have time to figure out how to 'de-gnome' my
> interface.
> When I switch to say, windowmaker that damn Gnome taskbar is still there
> with WM laid over top of it. The memory? I'm not making it up. Unless free
> is lying, I'm swapping when I never used to. Of course I think RH has
> installed every daemon known to nix, of which 50% I don't need. I haven't
> gotten around to trimming the modules.
I have tried Mandrake, too. When you choose Sawmill at login there,
you only get the bare screen and the xterm. No toolbars or anything.
Very nice for monitoring work.
--
You have changed the signature included in your e-mail.
For these changes to take effect, you must restart your computer!
Do you wish to restart your computer now?
[YES] [NO]
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sat, 15 Jul 2000 12:52:08 -0500
On Sat, 15 Jul 2000 13:30:14 -0400, [EMAIL PROTECTED] (tinman) wrote:
>In article
><[EMAIL PROTECTED]>,
>[EMAIL PROTECTED] wrote:
>
>> On Fri, 14 Jul 2000 15:44:40 GMT, Passenger Pigeon
>> <[EMAIL PROTECTED]> wrote:
>>
>> >In article
>> ><[EMAIL PROTECTED]>,
>> >[EMAIL PROTECTED] wrote:
>> >
>> >> On Fri, 14 Jul 2000 15:11:25 GMT, Passenger Pigeon
>> >> <[EMAIL PROTECTED]> wrote:
>> >>
>> >> >In article
>> >> ><[EMAIL PROTECTED]>,
>> >> >[EMAIL PROTECTED] wrote:
>> >> >
>> >> >> Poor Mac. When did the Mac get 256k or 512k?
>> >> >
>> >> >Now, I know this is a difficult logical leap, but somehow I just assumed
>> >> >that the Mac 512k had, oh, somewhere in the ballpark of 512k of memory.
>> >>
>> >> No shit. But when?
>> >
>> >http://search.metacrawler.com/crawler?general=mac+models&method=0&redirec
>> >t=web&rpp=20&hpe=10®ion=0&timeout=0&sort=0&format=beta99&theme=classic
>> >&refer=mc-search
>>
>> When? Can you give me a year? That link leads to many other links,
>> having nothing to do with the subject at hand. I'm sure you know -
>> just tell us.
>
>
>The Mac128 came out 1/24/1984.
>The Mac512 came out 9/10/1984.
In other words, both came out _well_ before the Amiga 1000, with 256k
of RAM. Thanks for proving my point.... (no offense to you; the
original poster who answered "The Amigas had more RAM" or somesuch was
wrong, though)
------------------------------
From: "Christopher Smith" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Linsux as a desktop platform
Date: Sun, 16 Jul 2000 04:00:55 +1000
<[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Sat, 15 Jul 2000 13:30:14 -0400, [EMAIL PROTECTED] (tinman) wrote:
>
> >The Mac128 came out 1/24/1984.
> >The Mac512 came out 9/10/1984.
>
> In other words, both came out _well_ before the Amiga 1000, with 256k
> of RAM. Thanks for proving my point.... (no offense to you; the
> original poster who answered "The Amigas had more RAM" or somesuch was
> wrong, though)
Yes, but the point is when they were _designed_, not released.
------------------------------
** 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
******************************
