Linux-Advocacy Digest #548, Volume #26 Tue, 16 May 00 22:13:09 EDT
Contents:
Re: Here is the solution (Christopher Browne)
Re: Is the PC era over? (Christopher Browne)
Re: progamming models, unix vs Windows (Bob Hauck)
Re: Ten Reasons Why Linux Sucks (JEDIDIAH)
Penguin Divorce ("ax")
Re: Microsoft finally gets the idea... almost ("Colin R. Day")
Re: Desktop use, office apps
Re: Beowulf (Gary Hallock)
Re: Here is the solution (Marty)
Re: progamming models, unix vs Windows (Gary Hallock)
Re: Why only Microsoft should be allowed to create software (Timberwoof)
Re: Why only Microsoft should be allowed to create software (Timberwoof)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To:
comp.sys.mac.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy
Subject: Re: Here is the solution
Reply-To: [EMAIL PROTECTED]
Date: Wed, 17 May 2000 00:27:46 GMT
Centuries ago, Nostradamus foresaw a time when Leslie Mikesell would say:
>In article <03FT4.171$[EMAIL PROTECTED]>,
>By the way, does anyone know where the SMB protocol really originated?
>It seems identical to the original IBM PC-network protocol back
>in the DOS days. Was that from IBM, MS, or did someone else invent
>it?
The salient question would be of who paid for the first LAN Manager
implementation.
<http://www.performancecomputing.com/unixintegration/9608/9608u1.htm>
indicates:
"... IBM and Microsoft developed a family of proprietary protocols to
support networking on the PC. On top of the naming and transport
functions of NetBEUI, IBM and Microsoft built a Server Message Block
(SMB) protocol for networking. In the mid-1980s, IBM and Microsoft
used SMBs to develop network software for the PC that provided file
and printer sharing."
This implicates "joint blame" for SMB to IBM and Microsoft.
Note that this took place at around the same time that IBM and
Microsoft were good friends, when IBM was paying Microsoft to do
development on OS/2.
--
"HE'S AHAB, boy; and Ahab of old, thou knowest, was a crowned king!"
-- /Moby-Dick/, Ch 16
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/lsf.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Is the PC era over?
Reply-To: [EMAIL PROTECTED]
Date: Wed, 17 May 2000 00:27:47 GMT
Centuries ago, Nostradamus foresaw a time when Tim Tyler would say:
>I think Donal K. Fellows <[EMAIL PROTECTED]> wrote:
>: In comp.os.ms-windows.nt.advocacy Tim Tyler <[EMAIL PROTECTED]>
>:> With no crappy x86 instruction set, the rest of the hardware might
>:> have some chance to overcome the traditional drawbacks of the
>:> ancient IBM-clone.
>:
>: The bus architecture still leaves a lot to be desired; though PCI is
>: much better than ISA, that isn't saying a right lot! For
>: everything else, there are either reasonable solutions about now
>: [...]
>
>That's not my impression. I'm no expert in this area, but ISTM that
>peripherals still need to fight over interrupts, the memory is still
>treated strangely - and support for decent parallel processing sucks.
>
>I hate the way expansion cards, go straight into the motherboard, and
>require disassembly to modify - the whole design seems dreadful.
>It looks like something from before the age of consumer electronics -
>which is, in fact, what it is.
What "the world needs," perhaps, is to have I/O managed via a set of
asynchronous buses.
Whether it's the "perfect answer," solving the world's problems, or
not, <http://www.pcaccelerate.com/News/SerialATA/serialata.html>
Serial ATA looks to have the _basis_ for the properties that would be
desirable, moving forward.
--> Attaching devices using [reasonably fast] serial buses means that
they aren't tied so strongly to the CPU or memory buses.
--> The use of a serial bus, with an asynchronous message-oriented
protocol similar to SCSI command queue tagging, would permit a
considerable degree of parallelism.
--> It would appear natural to trade the usual 50-odd pins being used
to connect in an IDE or SCSI bus for half a dozen 8-pin
connections similar to serial ATA or Firewire, _each_ capable of
supporting multiple drives.
--> One slick thing about SCSI as well as serial UARTs has been that a
few bytes worth of buffer, at the bus level, can make an _immense_
difference in terms of performance.
Toss the critical 64 _bytes_ (that's bytes, not "gigabytes") of
memory onto the UART, and CPU loading would drop _massively_.
This was _crucial_ back in the days of the 16450 UARTs, when
without the extra buffering, an 80386 CPU would get _killed_ if
you tried to drive a modem faster than 4800 bps.
Similar is true for SCSI host adaptors; a few bytes really helps.
<http://www.tdl.com/~netex/controllers/controllers.html> is
suggestive that being able to queue 1500 commands is the "state of
the art" today; that probably represents about 8K or 16K of
memory. I recall when it was a big deal that Symbios/NCR chips
could support a 256 command queue. Command queueing is what
_STILL_ makes SCSI superior to IDE.
Make queueing of commands pervasive on a somewhat more generic
bus, and this could provide substantial I/O improvements even if
the bus is a bit slower than the latest variation on SCSI.
I don't think this is _happening_, mind you, but it's the direction
I'd sure like to see...
--
Jumping off a cliff doesn't kill you! It's only when you hit the
ground...
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/lsf.html>
------------------------------
From: [EMAIL PROTECTED] (Bob Hauck)
Subject: Re: progamming models, unix vs Windows
Reply-To: bobh{at}haucks{dot}org
Date: Wed, 17 May 2000 00:27:21 GMT
On Tue, 16 May 2000 06:26:17 GMT, [EMAIL PROTECTED] <[EMAIL PROTECTED]>
wrote:
>> > Segment FFFF has absolutely nothing to do with A20. You can
>> > construct addresses in segment FFFF which lie below 1M.
This is correct of course.
>> > The A20 problem is exactly what it says: the A20 line on the bus. It
>> > has absolutely nothing to do with the internal segmentation of the
>> > processor.
Yes, it is the physical address that's important. On the 286 and higher,
A20 needs to be disabled (held at 0) to run some real-mode apps. These
apps assume that physical addresses "wrap around" to low memory. This was
the case on the 8086, which had only address lines A0-A19. The A20 gate
simulates this condition.
OTOH, this gate must be set to enable A20 for protected mode apps to work
correctly (which includes DOS "memory extender" software).
--
-| Bob Hauck
-| To Whom You Are Speaking
-| http://www.bobh.org/
------------------------------
From: [EMAIL PROTECTED] (JEDIDIAH)
Subject: Re: Ten Reasons Why Linux Sucks
Date: Wed, 17 May 2000 00:36:49 GMT
On Tue, 16 May 2000 20:22:58 -0400, mlw <[EMAIL PROTECTED]> wrote:
>Charles M wrote:
>>
>> On Mon, 15 May 2000 20:50:21 -0400, mlw <[EMAIL PROTECTED]> wrote:
>>
>> >Syphon wrote:
>> >>
>> <snip>
>> >
>> >>
>> >> 6. Supported printers? Damm better be a Postscript printer, linsux
>> >> seems to be the only folks using these printers these days....
>> >> Otherwise you will be burdened by some filter that a pimple faced geek
>> >> dreamed up that won't utilize 10 percent of your printers
>> >> capabilities.
>> >
>> >I have an HP 870cse, it is not a post script printer, but it works fin
>> >under Linux. Your claim is baseless.
>> >
>> >>
>>
>> A bit OT but ,how does the HP 870ce do with color graphics like
>> reproducing photos? I'll have to admit that I think he had a valid
>> point here (even if trolling) in that I've never gotton the same
>> level of color reproduction out of the ghostscript drivers on Linux
>> that I get out of the manufacturers native Windows drivers.
>> I'm using an Epson Stylus color 900 right now, for instance ,and I've
>> tried other "near photgraphic" quality printers under Linux and for
>> photos I can't find any setting that gives me more than mediocre
>> quality, even though these printers produce very good results under
>> Windows. Any hints on how to achieve this?
>
>The OEM printer drivers under Windows are actually very specialized
>rasterizing programs designed especially for the hardware. For me, photo
>printing is not an issue. I am an amateur photographer and find even the
>best home/office printer sorely lacking (yes, even under Windows), I
>don't consider bad and "less bad" really selling points. I can produce
>and edit photo quality images, print them so see what the print will
>eventually look like. Then I take it to a service house with some real
>imaging hardware. (These do NOT run Windows BTW.)
>
>I know this is sort of dodging the point, but, it is a real issue. The
>printers are getting better of course, but they are still inadequate for
>any real imaging work. To argue that this environment is better than
>that, because its printing of pictures is less inadequate with existing
>hardware, is a ridiculous position.
Not at all. This is consumer grade hardware and the differences
in photo output quality are quite significant. Better rendering
of color for print purposes is sorely needed in Linux. However,
it's not something that needs to be addressed on a per printer
basis.
--
In what language does 'open' mean 'execute the evil contents of' |||
a document? --Les Mikesell / | \
Need sane PPP docs? Try penguin.lvcm.com.
------------------------------
From: "ax" <[EMAIL PROTECTED]>
Subject: Penguin Divorce
Date: Wed, 17 May 2000 00:43:52 GMT
The proposed marriage between Corel and Inprise is breaking up.
Every divorce is ended with tears. But this one is ended with cheers.
How funny that is? I wonder how this pair of sophisticated penguins
could ever fall in love with each other at the first place? They are
pretty romantic!
------------------------------
From: "Colin R. Day" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Microsoft finally gets the idea... almost
Date: Wed, 17 May 2000 00:47:59 +0000
"R.E.Ballard ( Rex Ballard )" wrote:
<snip>
>
> > I think people should be able to /view/ executable
> > content (as far as it is viewable)
>
> bad idea. Better to detach it to a file, examine it with a
> debugger or binary file examiner, or the appropriate text editor,
> than to hap-hazardly begin executing a script that may do anything
He said VIEW, not execute. Just open the script in an editor.
>
> from ship your password to your competitors all the way to
> reformatting your hard drive.
>
> > from the mail client, but not launch it from there.
>
> Therin lies the rub. One of the advantages to object oriented
> systems is that the reader is associated to the content. But
> since the nature of a viewer is to cause a file to be displayed,
> one must assume that a script file will be displayed by executing
> the script.
>
Why couldn't it be displayed in a text editor?
Colin Day
------------------------------
From: [EMAIL PROTECTED] ()
Subject: Re: Desktop use, office apps
Reply-To: [EMAIL PROTECTED]
Date: Wed, 17 May 2000 01:19:07 GMT
On Tue, 16 May 2000 23:02:53 GMT, JEDIDIAH <[EMAIL PROTECTED]> wrote:
>On Tue, 16 May 2000 22:45:36 GMT, Raul Valero <[EMAIL PROTECTED]> wrote:
>>K6-2/350MHz, 256MB SDRAM PC-100, UDMA66 8GB HD, kernel 2.2.15 with VIA Bus
>>Master
>>XFree86 4.0, any window manager or desktop environment (often iceWMm,
>>Enlightenment or KDE)
>>
>>The whole Staroffice and an open document under the word processor takes
>>less than 1 minute by far.
>
> My K6-3/400 w/ 128M & UDMA 20G Deskstar is the same way. However,
> the same version of StarOffice under Win32 and a mere 64 loads
> slowly AND bogs the system down in the process.
>
> SO5 a memory hog. Whodathunkit? '-)
At least it doesn't have dancing paperclips, whole paragraphs that indent
when you indent the last line, or an insistance that the word following a
sentance be capitalized, no matter how many times you uncapitolize it.
Word is an abomination. They should have quit in '94.
------------------------------
Date: Tue, 16 May 2000 21:19:06 -0400
From: Gary Hallock <[EMAIL PROTECTED]>
Subject: Re: Beowulf
mlw wrote:
>
> Any good reason to choose PVM over MPI? (I am leaning toward MPI.)
I would definitely go with MPI. Like PVM, it is readily available on
lots of platforms. But there are platforms that have highly optimized
MPI. If you decide someday to run on an IBM SP machine with a high speed
switch, your code will really fly without change.
Gary
------------------------------
From: Marty <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy,comp.sys.mac.advocacy
Subject: Re: Here is the solution
Date: Wed, 17 May 2000 01:24:40 GMT
[EMAIL PROTECTED] wrote:
>
> In <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Bob
>Hauck) writes:
> |On Tue, 16 May 2000 02:00:04 GMT, [EMAIL PROTECTED] <[EMAIL PROTECTED]>
> |wrote:
> |
> ||This is an _extremely_ close minded view as you do not consider any
> ||systems besides Unix, Mac, and Windows. Did you even know that (GASP!)
> ||other systems actually exist?
> |
> |Actually, yes I do know that. VMS is so far from the right solution for a
> |home user that it never entered my mind. Similarly, I would not recommend
> |MVS or TOPS-20 or VxWorks for the home or small business user.
> |
> |
> ||The best OS for an OS/2 user to move to is VMS. Point by point:
> |
> |You have _got_ to be kidding. If you really believe that VMS is the right
> |system for someone coming from OS/2, then you are a lunatic.
> |
> | -| Bob Hauck
>
> Oddly enough, VMS is one of a short list of major OS's I haven't had
> a chance to play with; although, some folks I respect from the VM days
> thought that VMS was pretty good stuff...
>
> Isn't VMS a batch OS underneath, like MVS? I suppose you might replace
> some OS/2 server usage with that; but, what about the interactive stuff.
> I don't imagine VMS is exactly overflowing with word processors and
> stuff... Doesn't seem a likely choice to me.
It has an X interface and a very nice help system which blows the doors off of
the standard *nix "man". However, the thought of using it as a replacement
for a PC is silly, as the hardware prices are in a different league.
------------------------------
Date: Tue, 16 May 2000 21:37:15 -0400
From: Gary Hallock <[EMAIL PROTECTED]>
Subject: Re: progamming models, unix vs Windows
David Cueto wrote:
> To talk about inherited things, take a lot at your GNU/Linux box,
> trying to implement an ancient Unix model that even proved good
> for a lot of things, has a lot of supidities coming from the age of
> stone. Unix was designed when current hardware, networks and
> software did not exist, so adapting itself to nowadays is as difficult
> as for Windows (9x) leaving drive letters.
Ok, I just looked at my GNU/Linux box. The "ancient" hardware includes
an OSA network card and a bunch of IBM 3390 disk drives. Looks pretty
modern to me. Oh, I almost forgot that ancient CPU - a 12-way IBM
S/390 G6.
Gary
------------------------------
From: Timberwoof <[EMAIL PROTECTED]>
Crossposted-To:
comp.sys.mac.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy
Subject: Re: Why only Microsoft should be allowed to create software
Date: Tue, 16 May 2000 18:48:16 -0700
In article <CG9U4.12986$[EMAIL PROTECTED]>,
"Daniel Johnson" <[EMAIL PROTECTED]> wrote:
> "Timberwoof" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > In article <Ij0U4.286$[EMAIL PROTECTED]>, "Erik Funkenbusch"
> > <[EMAIL PROTECTED]> wrote:
> [snip]
> > > And what IIS hooks are there in the OS for applications to use, other
> > > than the documented ISAPI?
> >
> > Nobody knows. They're not documented!
>
> Then how do we know they exist at all?
Run an application in a debugger and have it flag you whenever it makes
calls into system DLLs. When it makes calls no one has ever heard about
or can find documentaiton for, you've got the evidence.
<snip>
> The things MS calls "common controls" shipped with Windows 95. There
> are a bunch of *other*, more pedestrian controls that shipped with
> Windows *1*.
>
> It isn't entirely obvious what is so uncommon about things like
> the button control. But there you are.
Standard buttons. Whoopeee. The Mac shiped with standard file-open,
file-save, and print dialogs from the start. Windows did not.
> > The Mac shipped with its Toolbox right from the start, and
> > they openly documented the calls for that toolbox right from the start.
>
> Not unlike Windows, that way. :D
Well, Windows had a good model to copy ... the Mac.
> [snip]
> > > No, if you break up the application and OS division, it means that
> > > 3rd
> > > party
> > > developers will no longer be able to take advantage of the
> > > Application
> > > division written code that MS rolls into the OS. This will put 3rd
> party
> > > developers at even more of a disadvantage.
> >
> > Funny ... no one at Adobe or Macromedia seems to be complaining about
> > the Mac OS's lack of support for applications.
>
> Perhaps they should be. MacOS's application support is decidedly
> inferior to Windows. Ferinstance, if Photoshop had been written
> for Windows 95 or NT in the first place, I doubt they'd have rolled their
> own VM implementation.
>
> Now, it's true that Win95 and NT didn't exist back then, and MacOS did-
> but it is the year 2000 now, and MacOS *still* doesn't have a VM
> implementation adequate to Photoshop's needs.
Windows VM doesn't meet many applications' needs. Smartheap is a set of
libraries that takes care of all sorts of memory allocaiton and
deallocation details for you that Windows dows not do that the Mac OS
does.
> If they aren't complaining, it's probably because they don't think the
> Mac
> will survive, and so feel its not worth bellyaching about.
This simply shows that you don't know what you're talking about. Both
Macromedia and Adobe get about half of their income from Macintosh
applications. (And which major software company did you work for as a
software engineer?)
--
Timberwoof
Chief Perpetrator, Infernosoft: Putting the No in Innovation.
(timberwoof at infernosoft dot com -- www dot infernosoft dot com)
------------------------------
From: Timberwoof <[EMAIL PROTECTED]>
Crossposted-To:
comp.sys.mac.advocacy,comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy
Subject: Re: Why only Microsoft should be allowed to create software
Date: Tue, 16 May 2000 18:48:15 -0700
In article <8fs6jg$2lac$[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Leslie
Mikesell) wrote:
> In article <diaU4.320$[EMAIL PROTECTED]>,
> Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
>
> >If I switch to MacOS, either I'm going to throw away all my old PC's and
> >buy
> >new Mac's, or I'm going to add new Mac's as new people come in and throw
> >away PC's as they break (replacing them with Mac's). That's not going
> >to be
> >a cost difference any more than buying new Ford's to replace aging
> >chevy's.
>
> So, on a new machine the place you really lose money is where the
> cost of an OS you don't want and won't use is bundled with
> the hardware.
>
> Les Mikesell
> [EMAIL PROTECTED]
Oh, you mean like how Windows is bundled with PCs?
--
Timberwoof
Chief Perpetrator, Infernosoft: Putting the No in Innovation.
(timberwoof at infernosoft dot com -- www dot infernosoft dot com)
------------------------------
** 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
******************************
