Linux so far no software hare (fwd)
CLIFFORD ILKAY
clifford_ilkay-biY6FKoJMRdBDgjK7y7TUQ at public.gmane.org
Tue Jan 27 16:41:23 UTC 2004
At 08:20 27/01/2004 -0500, cbbrowne-HInyCGIudOg at public.gmane.org wrote:
> > I just sent this to the Star. I sent it as an article to the Business
> > section. I don't really know the best place. Perhaps their
> > Ombudsman.
>
>[Much elided that I agree with; it is _very_ well written, Hugh!]
Agreed.
> > MS Office itself is unlikely to be more stable on Linux than on MS
> > Windows.
>
>I'll take a _bit_ of issue with this, as there have been known to be
>surprises.
>
>I have seen several reports from people I trust well with the rather
>surprising conclusion that running Windows atop VMWare atop Linux leads
>to an environment noticeably more stable than running Windows
>"natively."
I do both, run Windows natively and on VMware that is, and have not found
any difference in stability. All modern implementations of Windows, i.e.
NT, 2000, and XP, have actually been quite stable for me, for some narrow
meaning of the word "stable". If the ability to continue running without
blue screening, as the Win9x series is wont to do is a measure, then yes,
the aforementioned are much more stable. BSODs on NT, 2000, or XP are not
the norm, as some Linux/Mac OS bigots would like others to believe. What I
have found, however, is that all three of these operating systems suffer
from lousy memory management. This may have more to do with the
applications that I run on them, e.g. Eudora, Internet Explorer, than the
underlying OS but I find that after a few days, in order to maintain some
reasonable semblance of performance, I have to do a restart. Otherwise,
memory usage balloons out to grotesque levels and Windows starts swapping
heavily. Windows hitting swap is not a pretty sight. More RAM only delays
the inevitable.
A difference I have noticed with running Windows on VMware vs. running it
natively on the same hardware is that I/O is considerably slower running on
VMware, which I suppose is understandable.
>The apparent explanation is that the "simulated PC" that VMWare provides
>has fewer "jagged bits of functionality" coming in from the hardware
>side that can 'break down.' In effect, the "virtual machine" that
>their product emulates is better, for the purpose of running Windows,
>than a real one.
>
>That is quite different from what I would be inclined to expect; what I
>would _expect_ is that VMWare, being a "simulator," with all sorts of
>"kludges," would be considerably less reliable than a "real machine."
>
>An implementation of MS Office that uses the boundaries of Unix
>processes and directories and permissions to enclose its behaviour might
>conceivably be more stable than the existing editions that have
>generally been pretty unconstrained in what they can do to one's
>computer system.
>
>Of course, the opposite can be true; witness the reports of the hideous
>implementation of Internet Explorer on Solaris and HP/UX...
>
>But it is good to keep in mind that Microsoft has implemented MS Office
>atop "Unix" in the form of the latest edition for MacOS X. After doing
>"it" (porting to Unix) once, Microsoft probably has enough experience to
>be able to choose what degree of stability they want unleashed...
My understanding is that Office X, the OS X version of office, was NOT a
port but a clean slate design. My friends who run it love it and they are
not big MS fans so clearly, MS can write solid code if they want to. Word 6
probably caused more defections from the Mac OS to Windows in the mid '90s
than just about any other single factor, which could have been a deliberate
strategy if it were not simple incompetence. Word 6 *was* a port, and an
ugly one at that. It was the biggest piece of junk and crashed early and
often. I remember the arcane incantations we had to mumble and the ritual
sacrifices that we had to make just to have System 7.6x stable enough that
it would "only" freeze couple of times a day. Turn virtual memory off, turn
virtual memory on. RAM Doubler, Speed Doubler, disable this init, enable
that one, load inits in a different order. Not too many fonts, not enough
fonts, not the right fonts... blah!
I found the comment in the Star article about crashes on opening MS Office
documents to be amusing because the only time that I have ever experienced
that is with MS Word (XP) opening a file created using the *same* version
and instance of MS Word. I could get Word to GPF reliably on this one
particular file but OpenOffice on Linux was happy to open it, modify it,
and save it back in Word format.
Password protected MS Office documents are a problem even with other
versions of the same application, much less OpenOffice. We have a client
that has stopped using OpenOffice and is standardizing on Office XP mainly
because of this reason. They had lots of Office 97 and Office 2000
documents and sharing between the two versions did not seem to be a
problem. Enter OpenOffice, and it could not unlock password protected
Office documents. So, they "upgraded" to Offfice XP and lo and behold, it
could not unlock their existing password protected documents either. I
questioned why they needed to password protect documents anyway since all
that I have ever read about this practice suggests that it is akin to
leaving one's house key under the door mat and it seems like there are an
inordinate number of problems with file corruption with this practice, as
evidenced by tales of woe and the number of utilities to "recover" MS
Office files on the Internet. It was their way of protecting sensitive
documents sent by email from prying eyes instead of using PGP. I've come to
the conclusion that email encryption is too complicated for the average
user so it does not get done and people resort to hacks like password
protecting documents or zip files.
Regards,
Clifford Ilkay
Dinamis Corporation
3266 Yonge Street, Suite 1419
Toronto, Ontario
Canada M4N 3P6
Tel: 416-410-3326
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
More information about the Legacy
mailing list