BH came forth with the following pronouncement in alt.os.windows-xp on
18 Oct 2001 11:55:31 -0700:
>I can blame the XP OS. After I could not get some hardware/software to
>work such as: Scanners, Wireless PCMCIA cards, and games.
Which still doesn't tell us dick, and consequently does not allow us
to say *where* the problem lies...
That games targetted at Win9x *and often specifically exclude NT
kernel O/Ses* can cause problems is not exactly a revelation. Games
programmers and developers in particular are notorious for engaging in
all sorts of tricks that they really shouldn't in order to maximise
performance and do 'interesting' things, very often blurring the line
between what should be handed off to the O/S and what should be
handled by an application - if not, on occassion, slamming straight
through that line. The bug lies in the 9x O/Ses, because they should
have been designed such that this *can not happen*.
Does this mean that XP is not ready for release to the public? Of
course not - it means 9x should never have been released to the public
as it is a weakly designed, inherently insecure and flawed O/S.
Drivers are a more complex issue, but the same ideas can still apply.
> I formatted
>the drive and installed Win ME and everything works fine. I think I
>covered my bases.
Products targetted at 9x platforms seem to function properly on a 9x
platform. What an astounding revelation.
>Oh and Win ME does not ask me to send a mini-dump
>after it crashes to Microsoft for review. That was a funny one. Just
>proves I am a beta tester, and not getting paid to be one.
Every flavour of *nix allows kernel dumps et al when run-time and
system errors occur. Does this mean that they are beta products? Nope.
It simply allows users/admins et al to identify what went wrong and,
more importantly, why. With this information it is relatively trivial
to resolve the problem and prevent it occuring again.
So, why does M$ offer to analyse this data for you? I can only
speculate, but it is clearly apparent that there are two problems with
error-tracing and tracking in XP: (i) There are many users migrating
to XP who have no real idea as to how systems work and consequently
will have no ability whatsoever to analyse this information; (ii) XP
(like all M$ products) is a closed, proprietary technology so it is
much harder for all users to understand the intricacies of the O/Ses
functional processes without access to developer kits and appropriate
reference materials.
Why is this functionality there? Because things will go wrong. It
doesn't matter what you're using, in an entropic universe problems are
inevitable from time to time. This is exacerbated on IBM PC
compatibles for the simple reason that it is an open platform - it may
have pre-defined and supposedly agreed upon standards, but for many
systems engineers, designers, developers and manufacturers (all at
both hardware and software levels) standards are something that happen
to other people.
This means that whilst most people may not have a problem, there will
always be those that have a resolvable but convoluted issue and those
that have an unresolvable one.
With 9x O/Ses, however, attempting to identify the cause of a problem
- let alone resolve it - is about as entertaining as being buttfucked
by a freight train.
>Enjoy the slight headaches while trying to be productive with XP.
Oh i will, because i know that almost all problems i encounter will be
the result of things that i did, rather than the O/S being an
appallingly designed heap of garbage. And, on those occassions when i
have had a problem or those i have in future, i will actually learn
something in resolving it (like randomly disabling and enabling
functionality in gpedit.msc can cause real headaches ;-).).
> I
>will work with 98/ME/2000 until the service patch comes out, and I see
>the hardware manufactures bring out new XP drivers.
Fine. If it works for you then continue to use it. People should only
upgrade if they have a need to do so. But stop knocking a product when
you clearly have no idea what you are talking about.
Most of us, however, are totally unwilling to voluntarily go back to
an O/S that crashes for no sodding reason that has a stable timeframe
measured in hours if you're lucky and is about as secure as US
airports pre-September 11th...
> If those two
>things do not happen... Well then I will stick with 98/ME/2000 and
>invest my valuable time in Mandrake Linux
Much as i love the experience i have had with *nix environments
(having used everything from early Linux kernel builds to dumb
terminals with X-clients slaved to a SuperSPARC running Solaris for
everything from fiddling for amusement to compiling quantities of OO
code etc), the reality is that if you are having problems with 2k/XP
you are going to have a very difficult time with Mandrake if you wish
to really take advantage of its functionality and really make use of
those things that *nix environments still do far better than any
Windows build [1]. The learning curve can be far steeper, and the
online documentation can sometimes be woeful (i sometimes wonder why
they bothered including some of the man pages, although the fact that
Stallman seems to have co-authored half of them might explain their
crapness) - and that's without getting into the issue of
hardware/driver support for Linux.
Unfortunately, computers are not simple to use and never will be
unless you have well-developed analytical abilities and most people
don't. Even worse is the number of people who when confronted with
Lotus, Corel or whatever will go "But i only know how to use
Office!!!". Terrifying really.
[1] - I'm still not totally convinced that MS grasp how to properly
implement a server/client model or how to model a dynamic file system
that is transparent to clients on the server - it can be done but it
always seems to take far too much work. There are other things such as
the setting of group-specific policies and/or user-specific policies
that don't seem to implement as i was always taught that they should
be. I admit that i have little experience of MS Server-side builds so
may be missing on something that is relatively trivial to configure in
such an environment, but if so then that it creating a distinction
that shouldn't be there.
There's also issues related to granting group/user access or
privileges to trees, directories, files and processes which seems to
be overly complicated. Chmod (for example) is such a useful tool on a
*nix box, but i can't seem to find a comparably easy solution to do
this from a command line in XP. Maybe i'm missing something obvious...
And i want a tool as powerful and useful as grep ;-)
--
e-mail: g.tay...@freeuk.com
ICQ: 105762096
PGP key ID: 0x4B49334E
Fingerprint: F5D7 DEEA 72A4 6F93 0B8E 0D07 7070 8805 4B49 334E
May you die in bed at 95, shot by a jealous spouse.
|
