Linux-Advocacy Digest #683, Volume #26 Thu, 25 May 00 10:13:06 EDT
Contents:
Re: Why only Microsoft should be allowed to create software ("Erik Funkenbusch")
Re: Linux will never progress beyond geekdome (The Ghost In The Machine)
Re: W2K BSOD's documented *not* to be hardware (Was: lack of goals. ("Erik
Funkenbusch")
Re: Not so fast... (The Ghost In The Machine)
Re: Font deuglification ?? (Dave Rolfe)
----------------------------------------------------------------------------
From: "Erik Funkenbusch" <[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: Thu, 25 May 2000 09:01:15 -0500
Illya Vaes <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
> >>>DOS and Windows are OS's. They're not applications.
> >>>Windows cannot run without DOS, thus Windows and DOS are joined.
> >>(Then) Windows is not an OS. An OS runs without the help of another OS.
> >Gee, I guess DOS/VSE under VM isn't an OS then. I guess NeXT and MacOS X
> >aren't an OS then, since they rely on mach to function (actually there is
> >more than a passing similarity between the way mach and it's client OS's
> >run and the way DOS and Windows run.)
>
> *You* gave the criterion "X cannot run without Y, thus X and Y are
joined".
> Another strawman ofcourse...
What? I stated that X could not run without Y. You stated that this meant
that X wasn't an OS and I gave examples of other OS's that also required a
different OS to run. What's your point? How is it a strawman? It directly
refutes your statement.
> >>DOS is an OS, and DOS+Windows is an OS. Windows without DOS is not, it
> >>doesn't run (Windows NT excluded ofcourse).
> >And Darwin doesn't run without Mach either. mklinux doesn't run without
> >mach.
>
> And anything that uses Mach, if you want to look at it that way, just uses
the
> public API of Mach, not some internal data structures.
The source for MACH is available, thus everything can be a public API, even
code that is meant to be internal.
> >>Windows is as much an application as Borland C++ 3, Doom, GEOS, Dark
> >>Forces, ... All are "DOS Extenders" (which I don't have to explain to
> >>you).
> >>Windows just has a pretty complete (but not always consistent) API that
> >>"happens" to be used a lot (which is where their monopoly comes in).
> >Doom and Dark Forces arent dos extenders either,
> >they used DOS/4GW which was a DOS extender.
>
> And Schulman showed that Windows "used a DOS extender" too. The nature of
> these DOS extenders is such that they're included in the product
("binding")
> such that they themselves have become the DOS extender if they expose an
API
> to applications.
You're saying that Doom and Dark Forces expose API's to applications?
> >But, even so. There is a large difference between a DOS extender and
what
> >Windows does. DOS extenders merely call DOS from protected mode.
Windows
> >(while it provided an externder in the form of a DPMI server) actually
> >replaced many of the DOS services with protected mode versions.
>
> And providing a protected mode version of 'malloc' isn't "replacing DOS
> services with protected mode versions"?
> You are downplaying DOS extenders. Like their name implies they (can)
*extend*
> DOS; you're not extending anything (yet) by merely 'calling DOS from
protected
> mode' (whatever that may be).
Yes, the *CAN* extend DOS, and other than memory allocation, they generally
do not. For instance, they do not provide their own file system services,
Windows does.
> >>>An application can run on any system which provides the right API's
(such
> >>>as WINE), the OS cannot.
> >>Oh? So VMware makes several OS's not an OS.
> >VMWare does a whole lot more than provide an API. It has to completely
> >emulate non-virtualizeable portions of the CPU.
>
> Providing an API has nothing whatsoever to do with the complexity needed
to
> accomplish that fact.
> They provide an API 'equal' to the API a real PC box provides; if they've
done
> it successfully, the OS running on it sees no difference.
I'm sorry, but emulating a CPU is not an API.
> >I said "many of the DOS API's", to which you replied "That's *exactly*
what
> >a DOS extender does" but only gave one API. A DOS extender manages
memory
> >and provides real mode translation of DOS calls (such as disk access,
> >etc..). Windows, however, provides it's own disk management functions.
>
> So you want to argue that something becomes "more than a DOS extender"
because
> it happens to extend more of DOS than other DOS extenders that happened to
> also have been _marketed_ as such?
> I'd suggest a careful reread of Schulman's book.
> There is *no* _qualitative_ difference between Windows (excl. NT) and a
> general "DOS extender", only quantitative.
There's no qualitative difference between a chimpanzee and a human either.
We share almost the exact same DNA. There is however a huge quantitive
difference, using qualitative and quantitive as you do.
> If you want to argue something stops being a DOS extender when it
"replaces"
> more than xx% of DOS services in its own implementation, then fine, name
the
> number and come up with the number for Windows and other DOS extenders.
My argument isn't that it's not a DOS extender (I already said it provides
one). My argument is that it's a great deal *MORE* than that.
> Note, I'm not denying your tack of "Windows is an OS", but by the same
logic
> any DOS extender is an OS as well. If the latter isn't true according to
you,
> then Windows isn't an OS either. Pick one.
An OS is something that provides OS-like services. Not just memory
allocation, but file services, graphical input/output, device management,
etc.. Windows clearly provides those services, and for the most part
they're seperate implementations from the underlying DOS. Simple DOS
extenders do not.
> >>>Untrue. MS is a large company that many of it's clients expect certain
> >>Just keep beating that strawman!
> >>Next you'll tell us that MS has to guarantee AutoCAD to run on Windows.
> >Then why do they get blasted everytime an upgrade breaks someones
software?
>
> Because it's (usually) needless / on purpose and/or the broken software
used
> the API as documented???
But you claim that MS has no responsibility to maintain such compatibility.
Are you now changing that statement?
> >And why is that called an example of their Monopolistic behavior?
>
> Because it always happens to be competitor's products if there still is
> competition in that app area or their own product if they've already
> practically established a monopoly in the area of the app being broken??
MS apps are occasionally broken as well. But usually they aren't. MS tests
it's own apps against updates, it doesn't test every competitors app.
> >>>>The discussion is about their (alledged) right to *prevent* other
DOS's
> >>>>to run Windows (or feign incompatibility).
> >>>So, MS should just blindly run on whatever junk pretends it's MS-DOS
and
> >>>let the buyer beware?
> >>Yes.
> >What you fail to realize is that MS is the one that will be blamed for
> >problems caused by the other code.
>
> Get real. Applications get blamed for faults with Windows than the other
way
> around.
By people that know. Your average joe will blame MS though, rather than the
App.
> I'm sure the knee-jerk reaction of MS weenies to my NT machine freezing
> completely on a very regular basis is along the lines of "ah, but you
> shouldn't run Netscape but IE!". So much for a stable and robust OS...
Bugs happen.
> >>If you think not, please tell us why you expected DOS apps to run on
"OS"
> >>Windows and Windows NT (apps that just use the public DOS API)? Should
> >>they refuse to run on "whatever junk pretends it's MS-DOS"???
> >That's their right. In fact, many apps DO refuse to run in a Windows dos
> >box.
>
> (So much for your esteemed "MS backward compatibility")
> There's a difference between not running because you don't get enough base
> memory or are denied direct access to resources under control of the OS
and
> refusing to run because this isn't MS-DOS 6.21 but MS-DOS 6.22, DR-DOS
instead
> of MS-DOS or because you didn't find some signature in some internal data
of
> DOS.
> But I guess that's too subtle a distinction for you...
No, you missed the point. I'm talking about apps that don't even bother to
check if the services or memory they need is available. They just check for
running under Windows and fail.
An example of this is the many DirectX games that checked for NT and failed
to run or install if it was, assuming that NT didn't have DirectX. This
obviously failed when NT4 later got DirectX and with Windows 2000 which has
Direct3D.
> >>>You don't think MS has a responsibility to take a few
> >>>precautions against obvious stability issues?
> >>Certainly not when they're baloney or even counterfeit.
> >Novell admitted to a serious bug in DR-DOS that prevented it from running
> >under Standard mode Windows. Incompatibilities were not baloney or
> >counterfeit. They were real.
>
> MS seeking out DR-DOS / non-MS-DOS in order to give some bogus warning
message
> (and if they'd got away with it later no doubt refusing to run) doesn't
mean
> there are / can be no _real_ compatibility problems.
> So there was a real bug. It had nothing to do with the bogus message.
Whatever the purpose of the message is really irrelevant to this
conversation, which is simply that MS *DOES* have the right to warn people
of incompatibilities.
> >Many of the pre-windows 2000 aware games would not run on Windows 2000
> >because they checked for the existance of NT and refused to install or
run
> >if it was NT. They didn't bother to check if the API's they needed were
> >there, they just found it easier to assume NT was incompatible.
> >Are you suggesting that those people are part of a conspiracy as well?
>
> And if they used a public API for such boneheaded things to query the
version,
> then the API provider could give the user a way to specify for each
> application what version should be given to that app in order to
workaround
> the stupidity of the programmers of it (didn't W2K have something like
this?).
Yes, Win2k has something like this, but it's a hack.
> That's altogether different than poking in internal data structures for
one
> reason only: forcing people to run it only on your own API provider.
> And I'd guess "most" of those game companies didn't have a competing OS...
And what makes you think it was for one reason only? I've offered several
other viable reasons for it. I'm not saying that my reasons were the only
reasons either. Clearly, MS was also trying to protect it's monopoly, but
my point was that there *ARE* legitimate reasons to do this as well. Don't
take my arguments as being absolutes. Nothing is ever absolute (You'll
notice many of my arguments are arguing the absolute nature of the
statements made such as "for one reason only").
> >>>As an example, car companies frequenly provide circumstances in which
> >>>their
> >>>warranty will be voided if they do things which can cause the product
to
> >>>malfunction. Example, putting in unapproved motor oil. Since GM can't
> >>>modify their engines to prevent unapproved motor oil from being added
to
> >>>their cars, their only choice is to void the warranty.
> >>Does GM sell (inferior) motor oil?
> >How does that relate to the DOS/Window example?
>
> Hey, you gave the example...
And i'm asking you how it is relevant. What does inferior have to do with
anything?
> >>It's more like someone selling a car radio using a certain pin-out (API)
> >>that nevertheless automatically worsens its performance when it's put
into
> >>a pin-compatible bay not belonging to a car made by the same
manufacturer,
> >>while the car-radio is a monopoly product and the car is inferior ('by
> >>several measured criteria' to use a good-old-days-of-Mike-Timbol quote).
> >No. Again, Windows and DOS are both OS's that work together. A radio is
> >not a vehicle. It doesn't even provide vehicle-like services (which if
you
> >take the stand that Windows isn't an OS, you have to at least admit that
it
> >provides OS-like services).
>
> This is about APIs, not providing OS services. The car radio uses the BAPI
> (Bay API) and the car provides it. If the car manufacturer also provides
car
> radios and the car gives bogus warnings about other than their own car
radio
> even though they use the same API and you need a dependent car radio, then
you
> get near the case at hand.
No, the whole point is that both Windows and DOS are OS's or at least
OS-like. They provide similar services, and thus are roughly peers. A
radio is not a peer of a vehicle. A radio is an application, clear cut.
There is no such clear division between DOS and Windows.
> >>>MS could probably have refused technical support if your an DR-DOS, but
> >>>why make a customer angry?
> >>Screwing the customer out of money that needn't be spent (for
'upgrading'
> >>from DR-DOS to MS-DOS), *lying* to him and denying him to make that
choice
> >>for himself isn't going to make him angry????
> >There were legitimate and Novell acknowledge incompatibilities.
>
> And they had nothing to do with the bogus message. That would still have
come
> after Novell had fixed that.
As I said, MS has a legitmate right to limit it's products to work with the
products it knows will work. It doesn't matter if there are actual
incompatibilities or not.
> >Well, if that was it's intention, YOU should seriously consider your use
of
> >the language.
>
> I guess you still deny any possibility of you putting things "wrong".
> Too bad, your problem.
You're avoiding the statement. Your response, as you claim it was intended,
did not convey that message.
> >I doubt anyone else saw the phrase your way.
>
> I doubt that. We'll let the reader decide, which is just what I mean:
> regardless of how *you* say something was meant, the reader decides how it
> comes across.
Which is why you and others statements about what I intended with mine are
hypocritical.
> >>You tried to throw Joseph on a dead trail by starting about the
marketing
> >>of Win3.0/1, while the discussion uptil then had been about (the right
of)
> >>MS putting in tests in Windows specifically meant to single out
> >>non-approved DOS versions (that could otherwise be perfectly compatible
> >>products).
> >How a product is marketed is very much apropos to the intended purpose of
> >the product, which was to run on MS-DOS and PC-DOS.
>
> Rubbish.
> You market something that is ready, and if market conditions change (yeah
> right, in a monopoly) you can/must change it quickly.
> How MS markets Windows only shows how they *want* you to look at things,
it
> has no direct connection to reality. They *wanted* you to think there
would be
> incompatiblities, regardless of any factual incompatibilities.
> If you want to make 'marketing' leading in a discussion, you give liars a
> permit to say whatever they want.
> Hmm... yeah, I guess you *do* want to do that, since you've consistently
> defended everything MS did (even though you usually lag their 'position' a
few
> months)...
I'm not defending WHY they did them per se. I'm only saying that they have
a right to do them for legitimate reasons. I'm not saying that all their
reasons were legitimate, but I am arguing that not all their reasons were
nefarious in nature. I'm arguing the middle ground, which is that there are
neither all good reasons or all bad reasons, it's a mixture.
> >If MS marketed the product as an add-on to DR-DOS and then caused
> >deliberate failures, that would be one thing. But it wasn't marketed for
> >that, thus the buyers use of an unauthorized co-package is not MS's
fault.
>
> A "co-package"?? What a load of crap!!!
> You sure you're not a lawyer for MS (puke)???
co-package was the only way I could phrase it. What would you call darwin
and mach? I call them co-packages.
> >>Which has nothing to do with the public face they happened to have put
on
> >>at the same time. If you were to do that, would you say "we're singling
> >>out our competitor's product with this" or try to side-step it by
> >>lyin...eh... marketing?
> >MS's handling of the situation was poor in many ways.
>
> You mean they got caught (yet) again?
No, I mean their handling of it was poor. Rather than discuss legitimate
reasons for what they did, they choose instead to try and wash it all away.
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Subject: Re: Linux will never progress beyond geekdome
Date: Thu, 25 May 2000 13:52:25 GMT
In comp.os.linux.advocacy, win4win <[EMAIL PROTECTED]>
wrote on Thu, 25 May 2000 02:11:12 GMT
<[EMAIL PROTECTED]>:
>Sorry Unix/Linux geeks.. but Windows Wins.. I just installed Red Splat
>Linux and really had to brush the dust off the Unix memories to get it
>running. I'm so sure that your average user can wade through a Linux
>install and deal with all those Unix-ie messages! Not. Windows has
>NOTHING to fear until Linux can overcome its Unix-ness.
Well, all I can say to this is that I had occasion to install
Red Hat 6.2 on a malfunctioning Windows NT machine (it wouldn't
see the network and I couldn't log in) and found two minor glitches.
[1] Yes, it couldn't see the Windows network.
[2] No, it wasn't because of the hardware.
Oh, wait, that was *before* install. :-)
Once I figured out that the network was actually *connected*
(we use DHCP), I merely had to find an unoccupied IP using ping,
"borrowed" that for awhile, and started the download.
The download itself was smooth as silk (yours truly did have some issues
with the partitioning software, mostly because the first time I forgot
a '0' and RedHat complained that the root partition was too small,
and the second time apparently it ran into the end of the physical
disk and gave a 'bad parameter' error or something; I had to back off
1 megabyte on the /home partition, big deal); there was one
very minor documentation glitch (RedHat should somehow mention that what
is desired at the FTP prompt is the directory containing 'RedHat';
they instead use something like 'Red Hat System'), and a problem with
the mouse install (I'm going to have to work on that one) and X
(mostly because of the mouse).
Now, granted, I'm an expert at this; I've been installing systems
(non-professionally) at home for awhile: SLS, Slackware, RedHat 4,
RedHat 5, RedHat 6. One can interpret this two ways:
[1] That I'm required to install Linux more than once.
[2] That I can do it.
Now, [1] may be true; one of the things about Linux is its mutability.
RedHat 6.2 is a far cry from Linux's SLS days; it's slick, easy to
get to, and can pull itself off the net (one of my first downloads
was to 30 floppies for Slackware -- that worked reasonably well,
even back then). RedHat also has a "text" mode on its boot disk
for those who can't stand the GUI. (I used it mostly because I
didn't know what device and mouse driver I had.)
At my prior employment, I tried Debian; in some ways, Debian is even
slicker. However, I use RedHat because it's what I use at home.
I also note that FreeBSD can pull itself off the 'Net as well,
if my understanding is correct; this capability is not limited
to Linux.
In all fairness, Windows can also pull patches of itself off the
'Net, and any installer using InstallShield can apparently walk
the user through the steps required to install (sometimes incompatible)
DLLs that a program requires. Sometimes the user is required to
use a two-step process (e.g., for IE 5), but it's easy enough.
[2] apparently is an issue with Windows. It shouldn't be, obviously!
But a large number of Windows machines are preinstalled; the average
Joe probably would have troubles installing all of the free software
that comes with said machine.
Linux has a bit to go before the true neophyte -- a total newbie to
the world of computers who wouldn't know a mouse from a rat -- can hook
up a machine to the Internet, download some software, and have himself
a new OS. (This is probably one of the reasons why iMac sells so
well; just plug it in and go.)
But it's much farther along than Windows, at least in terms of 'Net
installs. In this case, Windows' propriety is actually a hindrance.
I'll merely mention also the free goodies such as GCC, TeX, ghostscript,
gnuplot, and xanim.
>Phtttt.
Excuse you. And next time open a window. :-)
--
[EMAIL PROTECTED] -- insert random misquote here
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: W2K BSOD's documented *not* to be hardware (Was: lack of goals.
Date: Thu, 25 May 2000 09:04:33 -0500
Adams Klaus-Georg <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Please try to get this into your head:
> If an application, however buggy it is, can crash an OS, it is the
> OSes fault.
I don't think anyone argues that having the ability to crash the OS is a
fault of the OS.
There is no argument that the application is faulty as well, though.
> Nowbody contends that the application has a problem. But if the OS
> crashes, it is buggy as well.
Yes, it is. But bugs exist in any OS.
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Subject: Re: Not so fast...
Date: Thu, 25 May 2000 14:00:24 GMT
In comp.os.linux.advocacy, Donal K. Fellows <[EMAIL PROTECTED]>
wrote on 24 May 2000 13:37:27 GMT <8gglun$r2r$[EMAIL PROTECTED]>:
>In article <[EMAIL PROTECTED]>,
>The Ghost In The Machine <[EMAIL PROTECTED]> wrote:
>> Yeah, but why did ILOVEYOU spread so fast?
>
>It used the most powerful forces in the universe.
>Ignorance and stupidity.
*grin*
Perhaps I should have asked "why is Windows so popular"? :-) :-)
[.sigsnip]
--
[EMAIL PROTECTED] -- I'd probably get the same answer :-)
------------------------------
From: Dave Rolfe <[EMAIL PROTECTED]>
Subject: Re: Font deuglification ??
Date: Thu, 25 May 2000 10:05:31 -0400
[EMAIL PROTECTED] wrote:
> And you never will find it. Linux looks ugly and like shit comapred to
> Windows.
>
> It is the dirty ugly secret, among others, of Linux.
>
> Do your eyes a favor and run an operating system that at least looks
> decent.
>
> I would suggest Mac.
I was in Sears yesterday and looked at the fonts on a Mac there. I was not
impressed. The fonts were very jaggy. Much worse than my linux machine. On my
linux machine I have a 19 inch display running at 1024x1280 with 100 dpi set
and the 100 pdi fonts first in the font list.
My complaint about linux fonts has to do with word processor support. On
windoze, if the fonts are defined to the system, then word processors know
about them and use them. Not so in linux. The word processors I have installed
can only access a small fraction of the installed fonts. In addition Netscape
cannot properly display web pages for this reason. I feel the web page display
problem is very serious and should be high on the todo list. My suspicion is
that these font problems have some kind of personality/political dimension that
I am unaware of. Otherwise this stuff would have been fixed long ago.
Dave
>
>
> On Thu, 25 May 2000 01:45:21 +1000, "Steve Budak"
> <[EMAIL PROTECTED]> wrote:
>
> >Howdy, I can't seem to find the howto for font deuglification in RedHat
> >Linux 6.2
> >I know I have it (or have seen it somewhere) but can't find it for the life
> >of me.
> >Anybody know a site where it might be located ?
> >I need to get the fonts to look a little better in Netscape and the like.
> >Thanks.
------------------------------
** 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
******************************
