Hardware Requirements for Internet PC

[BillW50]

In

In

snip
(Can't they be "told" it's a single-core processor they're running
on?) [Avira seems OK here.]
snip

Avira is the only AV that I can get to run on really old XP systems
with 512MB RAM or less and an older processor, without it bogging
down the entire system.

How do you mean, "tell" the program it's on a single core?

I monitor all of my processes all of the time for CPU usage. And
Avast 7 will run on anything from Windows 2000 and up. And AnVir Task
Manager says in the last hour, Avast hit 17% for a second and the
rest of the hour it was under 1%. As far as the average user is
concern, they couldn't even tell it was actually running.

Yes, but is that with 256 to 512 MB RAM as in my example? It's not
*only* CPU usage involved. As in the case of many apps, it's a combo
of both RAM amount and CPU. Avira seems to be the best one for very
low RAM systems, from my experience, anyway.... on a number of
clients' machines.

I know with Avast 4.x on older single core processors I've tried it,
no matter the RAM amount, definition updating would nearly max out
the CPU at the end of the updating. I can't say if there is an
improvement in that area with Avast 6 and 7, because I don't use it
on any machines. From what you've posted in your other replies, it's
been improved.

Well you got me there. I do have two Toshiba 2595XDVD laptops with only
192MB of RAM installed. But I don't use them for anything really. And I
don't update the AV or anything since I don't even use them connected to
anything anyway. But they should have Avast 4.8 still on them and I
never uninstalled.

Everything else of mine has 1GB or better of RAM. And I just checked
with AnVir Task Manager and Avast only taken a 1.4MB hit on the drive
for 3 seconds in the past hour. It did hit the drive again a few times
10 minutes later, but only about 200MB this time. It didn't do an
automatic update in this time so I'll keep an eye on it and see what
happens when it does.

 

[BillW50]

In

snip
(Can't they be "told" it's a single-core processor they're running
on?) [Avira seems OK here.]
snip

Avira is the only AV that I can get to run on really old XP systems
with 512MB RAM or less and an older processor, without it bogging
down the entire system.

(Good to know - I think - that I'm using a light one.)

How do you mean, "tell" the program it's on a single core?

I was referring to the suggestion someone made that most AV 'wares
these days assume they're on a multicore system, and cause heavy
load if run on a single-core system. I was wondering if they might
have a setting to prevent them trying to multiprocess. (Probably
not, as presumably it should be possible to do it - the detection of
whether multicore or not - automatically, so if they don't, they're
not going to bother.)

In Windows, you can use Task Manager and the "Affinity" setting,
to force an executable to stay on a particular core. On a multicore
system, that would effectively reduce the loading on the system.
On a single core system, Affinity is not going to help you as
you only have one core to begin with.

Task Manager also has "priority" settings, and you can experiment
with cranking down the priority setting. If two processes want to
run 100% on the same core, they split 50:50. If you drop the priority
of one process by one notch, then one process might get 75% while
the other gets 25%. So priority doesn't prevent one from running
entirely, it just changes the balance between the two. If you use
too extreme a setting, sometimes a potential side effect is a
deadlock in the system. So don't get carried away.

But an AV program, isn't going to tolerate being manipulated like
that. AV programs are pretty sensitive, and they have to be able
to defend themselves against any potential malware attack. Undoing
Task Manager changes, would be childs play for them. They're
armor plated.

Oh man! Manually adjusting the priority almost never works well. What
works better is software designed to automatically adjust the priorities
on the fly. As they automatically lower them if they are using too much
and bring them back up when they are using too little. Why Windows
doesn't have something like this built in, who knows.

 

[BillW50]

In

In

In message <[email protected]>, glee
message snip
(Can't they be "told" it's a single-core processor they're running
on?) [Avira seems OK here.]
snip

Avira is the only AV that I can get to run on really old XP systems
with 512MB RAM or less and an older processor, without it bogging
down the entire system.

(Good to know - I think - that I'm using a light one.)

How do you mean, "tell" the program it's on a single core?

I was referring to the suggestion someone made that most AV 'wares
these days assume they're on a multicore system, and cause heavy
load if run on a single-core system. I was wondering if they might
have a setting to prevent them trying to multiprocess. (Probably
not, as presumably it should be possible to do it - the detection of
whether multicore or not - automatically, so if they don't, they're
not going to bother.)

In Windows, you can use Task Manager and the "Affinity" setting,
to force an executable to stay on a particular core. On a multicore
system, that would effectively reduce the loading on the system.
On a single core system, Affinity is not going to help you as
you only have one core to begin with.

Task Manager also has "priority" settings, and you can experiment
with cranking down the priority setting. If two processes want to
run 100% on the same core, they split 50:50. If you drop the priority
of one process by one notch, then one process might get 75% while
the other gets 25%. So priority doesn't prevent one from running
entirely, it just changes the balance between the two. If you use
too extreme a setting, sometimes a potential side effect is a
deadlock in the system. So don't get carried away.

But an AV program, isn't going to tolerate being manipulated like
that. AV programs are pretty sensitive, and they have to be able
to defend themselves against any potential malware attack. Undoing
Task Manager changes, would be childs play for them. They're
armor plated.

Oh man! Manually adjusting the priority almost never works well. What
works better is software designed to automatically adjust the
priorities on the fly. As they automatically lower them if they are
using too much and bring them back up when they are using too little.
Why Windows doesn't have something like this built in, who knows.

What is funny is back in the Windows 3.x days, many claimed that
cooperative tasking was crap and preemptive tasking was great. And
Windows 3.1 didn't preemptive task anything but non-Windows applications
like DOS. I personally loved cooperative tasking and I thought it worked
really well. Although the tasking was dependent on all of the
applications to play nicely together. And the vast majority of them did.
The few ones that didn't well were not good programs IMHO and I wouldn't
bother using.

So now Windows 95 comes along and does preemptive tasking for 32 bit
Windows applications only. Many thought this was so great, I did not.
Now the applications had no say about how much CPU power they needed,
but now an OS who had no clue what the application was doing anyway did
say so when to start and stop (that is like having an idiot as a boss).
Now it is like you are working on something and you only needed a few
msec and you were done and the OS comes along and says you're done
before you finish. I never saw this as a good thing, despite my peers
back then.

Well, well, well... 17 years later with preemptive tasking and it just
isn't working so well. And all of those big promises how well life was
going to be and now it turns out that preemptive tasking just isn't
working well at all. Who would have guessed besides me who said it
wasn't going to work well? And where are those bozos who thought
preemptive tasking was our savior anyway?

Yes back in the cooperative tasking days, one bad apple that didn't play
well with others ruined it for everybody else. And now we have had 17
years of preemptive tasking and nothing changed. Still one bad apple
still ruins it for everybody. But it is worse now. It is like you are
about to take a bite of your favorite meal and get right to your mouth
and the OS says stop, it is somebody else's turn and you have to wait.
:-(

 

[glee]

snip
I was saying that if software is written to multiprocess, i. e. assume
that it has multiple cores to run on, then running it on a single core
will make it run inefficiently - and I was saying that ideally the
software ought to run in a different _way_ - one thing at a time - if
it is told (or, ideally, detects) that it is running on a single-core
system.
snip

I'm pretty sure no software writers and developers give a rat's nose if
you have an older single core system or less than 1GB RAM. If you do,
as far as they're concerned, you need to upgrade your hardware. They
simply don't write for older hardware, just like, for the most part,
they don't support older operating systems with their newer versions.

 

[BillW50]

In

I'm pretty sure no software writers and developers give a rat's nose
if you have an older single core system or less than 1GB RAM. If you
do, as far as they're concerned, you need to upgrade your hardware.
They simply don't write for older hardware, just like, for the most
part, they don't support older operating systems with their newer
versions.

True, but it bites them in the ass. As it is a two way street. I have
seen this over and over again. When programmers don't care about the
user's wishes, their income end up dropping. Only experienced
programmers realize this. The newer programmers don't get it and have to
learn the hard way. But they will blame it on something else of course
and might never learn.

Microsoft had played this game and balanced it very well all the way up
to XP. Then things changed. I personally think all of the older
programmers retired and there are only new programmers left who doesn't
know any better. So this is going to be a real test for Microsoft to see
how fast those newbies can learn this hard lesson.

It isn't all about Microsoft of course, that was just an example. There
are some software developers out there that still get it. Although I am
not sure if the future there will be any left.

 

[(PeteCresswell)]

Per BillW50:

When programmers don't care about the
user's wishes, their income end up dropping.

I had a really good (second-longest employed contractor in the
place) run at a major mutual fund based, I think, completely on
my attention to user's wishes. I'm not that great a developer,
but my motto the expression that a former co-worker coined: "I
don't sell programming. I sell happiness."

 

[BillW50]

In

Per BillW50:

I had a really good (second-longest employed contractor in the
place) run at a major mutual fund based, I think, completely on
my attention to user's wishes. I'm not that great a developer,
but my motto the expression that a former co-worker coined: "I
don't sell programming. I sell happiness."

Aww... I love it! ;-)

 

[Paul]

What is funny is back in the Windows 3.x days, many claimed that
cooperative tasking was crap and preemptive tasking was great. And
Windows 3.1 didn't preemptive task anything but non-Windows applications
like DOS. I personally loved cooperative tasking and I thought it worked
really well. Although the tasking was dependent on all of the
applications to play nicely together. And the vast majority of them did.
The few ones that didn't well were not good programs IMHO and I wouldn't
bother using.

So now Windows 95 comes along and does preemptive tasking for 32 bit
Windows applications only. Many thought this was so great, I did not.
Now the applications had no say about how much CPU power they needed,
but now an OS who had no clue what the application was doing anyway did
say so when to start and stop (that is like having an idiot as a boss).
Now it is like you are working on something and you only needed a few
msec and you were done and the OS comes along and says you're done
before you finish. I never saw this as a good thing, despite my peers
back then.

Well, well, well... 17 years later with preemptive tasking and it just
isn't working so well. And all of those big promises how well life was
going to be and now it turns out that preemptive tasking just isn't
working well at all. Who would have guessed besides me who said it
wasn't going to work well? And where are those bozos who thought
preemptive tasking was our savior anyway?

Yes back in the cooperative tasking days, one bad apple that didn't play
well with others ruined it for everybody else. And now we have had 17
years of preemptive tasking and nothing changed. Still one bad apple
still ruins it for everybody. But it is worse now. It is like you are
about to take a bite of your favorite meal and get right to your mouth
and the OS says stop, it is somebody else's turn and you have to wait.
:-(

Preemptive multitasking, is to *guarantee* correct operation, and
keep the kernel standing on its feet. They didn't do it for fun.

It was done, to allow building systems with "uptimes" longer than 24 hours.

If you compare MacOS 9 (cooperative) versus MacOSX (preemptive), the
difference is night and day. Twice a day, with the cooperative one,
I'd be down in Macsbug, trying to keep my computer running. The
preemptive OS is like a breath of fresh air by comparison. I never
have to rescue it.

Paul

 

[Stefan Patric]

In news:[email protected], Stefan Patric typed:

[Big snip]

I disagree that 900MHz isn't enough for any arbitrary video playback.
As

[snip]

My Asus EeePC 701/2 netbooks are underclocked to 633MHz. And they too
can keep up with arbitrary video playback without missing a beat under
Windows XP, even on an external monitor running 1440x900. Oddly
enough, Linux on the same machine can't even come close.

Which Linux distro? The one originally installed? Xandros, I think it
was. Awful.

I've got a EeePC 900 (900mHz Celeron, 1GB RAM, 4GB+16GB SSDs) on which
I installed Eeebuntu 3.x (an optimize version of Ubuntu for the EeePC)
wiping out the original Xandros, and it now plays any video, etc.
without problems. What a difference overall compared to Xandros.

Stef

Xandros, Ubuntu 8.10 netbook edition, Ubuntu 9.10 netbook edition, and
Puppy Linux. And I really liked Xandros, especially in easy mode which
boots in 20 seconds. Although the wireless to connect had taken an extra
minute. You could only use Firefox 2.0 tops with Xandros without
updating the kernel, and that makes Xandros unusable to me as is. As
Firefox 2.0 displays webpages worse than IE6 does.

Xandros was just too old, too, on my 900. And there were no newer
versions. That's one of the reasons why I replaced it with Eeebuntu
3.0. Unfortunately, Eeebuntu was based on Ubuntu 9.04, which went End of
Life last year, and along with that the repositories were removed. And
as Eeebuntu used them, not having its own . . . Well, that makes it all
the harder to keep it usable. (If only for security reasons, I should
upgrade Chrome and Firefox and Flash as they are all at least two years
old.)

I suppose in the future I'll be forced to install another OS on the old
EeePC. (Or retire it to my junk closet.) Don't know what it will be.
Eeebuntu was the only version of Linux I found--at the time--that was
specifically configured from the ground up for the EeePC, so everything
worked out-of-the-box. And it did. No glitches at all. No fixes
required. No tweaks needed. I was delightfully surprised.

Stef

 

[Searcher7]

Thanks everyone.

As I mentioned, surfing the internet is slow right from the beginning
of reformat and install, and it just gets worse.

When I said "Ctrl+Alt+Del" doesn't work I was thinking in terms of
turning my PC off *after* it freezes up. (But obviously in XP that is
not what it is supposed to do). Nevertheless, it does usually brings
up the Task Manager, but that may freeze also, necessitating an
unplug.

In the "Task Manager" Firefox appears to use the most memory at over
235k. (Perhaps because I have a lot of tabs open at the moment). I
have 31 items under "Processes" and when I refresh a window it will
shoot up to 100% under the "Performance" tab.

BTW. The system has Windows XP SP2.(I no longer bother with Windows
updates). I even uninstalled Google Chrome because of scripts which
kept trying to run even though I was only using Firefox.

As for hard drives, I've used different ones after each install, and
always have two in the system. A 20G and a 40G. (Not including the
500G USB drive). I'm only using about 1/3 of the space on my 40Gig "C"
drive. http://i290.photobucket.com/albums/ll257/Statenislander/Computer/SequoiaView.jpg

I'll ditch AVG and go with either Avast! again or Avira.

A pointer to how to turn off processes would also be appreciated. I
wish there was a way to avoid all scripts which are a main culprit in
freezing up my system. I only get individual options to stop the
script from running until the next one pops up.

Obviously this system cannot handle the internet efficiently anymore.
http://i290.photobucket.com/albums/ll257/Statenislander/Computer/BelarcAdvisor.jpg

I'm going to have to ditch this Dell XPS-Z Pentium III and build up
the hardware I already have, which include either a Dell 8400, HP
7955, or an ASUS A7V8X motherboard I can put in a case I have. All
have have a minimum of 2.8Ghz processors. I'll just have to make sure
the memory is maxed out or at least over 1G on whatever I system use,
and of course use a graphics card.

A dozen years ago internet speed on my 486 using dial-up beat the crap
out of the 1Ghz+ DSL systems I've used over the last decade. (Perhaps
I should dig out and stick to using Windows98 on one of these older
systems).

At the moment and as per "Speakeasy" my DSL speed is 1.57Mbps(Upload)
& .58Mbps(Download). According to Verizon I should get 3Mbps max, but
they conveniently don't say what the minimum is.

Thanks a lot.

Darren Harris
Staten Island, New York.

 

[glee]

Thanks everyone.

snip
In the "Task Manager" Firefox appears to use the most memory at over
235k. (Perhaps because I have a lot of tabs open at the moment). I
have 31 items under "Processes" and when I refresh a window it will
shoot up to 100% under the "Performance" tab.
snip

In Task Manager, while on the Processes tab, click the View menu> Select
Columns, put a check in the boxes for Virtual Memory Size and GDI
Objects, and click OK. You may find rather high Virtual Memory usage
for some apps (like Firefox) as well as for GDI Objects.... this will
tell you more about what is slowing it down that just looking at Mem
Usage.

snip
As for hard drives, I've used different ones after each install, and
always have two in the system. A 20G and a 40G. (Not including the
500G USB drive). I'm only using about 1/3 of the space on my 40Gig "C"
drive.
http://i290.photobucket.com/albums/ll257/Statenislander/Computer/SequoiaView.jpg
snip

....and what were the results of the Hitachi Drive Fitness Test (DFT) on
both internal drives, that I mentioned earlier?
http://www.hitachigst.com/support/downloads/#DFT

 

[BillW50]

In

Preemptive multitasking, is to *guarantee* correct operation, and
keep the kernel standing on its feet. They didn't do it for fun.

It was done, to allow building systems with "uptimes" longer than 24
hours.

If you compare MacOS 9 (cooperative) versus MacOSX (preemptive), the
difference is night and day. Twice a day, with the cooperative one,
I'd be down in Macsbug, trying to keep my computer running. The
preemptive OS is like a breath of fresh air by comparison. I never
have to rescue it.

Paul

I didn't see any of that with Windows. As far as stability went, each
newer Windows version was slightly better than the previous version. But
the change between cooperative vs. preemptive I didn't see anything
better as far as stability goes. Okay only in one area. And that was one
cooperative application could crash and take the rest of that session
with all of the other cooperative applications along with it. Although I
just stayed away from poorly written software and I didn't have that
problem.

And one big disadvantage with preemptive tasking that I found was that
multitasking and task switching got much slower compared to cooperative
tasking. Things were very snappy under cooperative tasking with 386 and
486 machines. Although when Windows started to use preemptive tasking,
those slower processors just couldn't cut it anymore.

Under today's machines with so much processor power, you probably
couldn't see the speed difference anymore. And my only big beef with
stability with the Windows 3.x and Windows 9x OS was the limit of
resource heaps which I never seemed to have enough of. Sure if you just
booted up, it probably would be fine for a few hours until the heaps got
low. And the quickest way to reset the heaps was to reboot.

 

[Paul]

In

I didn't see any of that with Windows. As far as stability went, each
newer Windows version was slightly better than the previous version. But
the change between cooperative vs. preemptive I didn't see anything
better as far as stability goes. Okay only in one area. And that was one
cooperative application could crash and take the rest of that session
with all of the other cooperative applications along with it. Although I
just stayed away from poorly written software and I didn't have that
problem.

And one big disadvantage with preemptive tasking that I found was that
multitasking and task switching got much slower compared to cooperative
tasking. Things were very snappy under cooperative tasking with 386 and
486 machines. Although when Windows started to use preemptive tasking,
those slower processors just couldn't cut it anymore.

Under today's machines with so much processor power, you probably
couldn't see the speed difference anymore. And my only big beef with
stability with the Windows 3.x and Windows 9x OS was the limit of
resource heaps which I never seemed to have enough of. Sure if you just
booted up, it probably would be fine for a few hours until the heaps got
low. And the quickest way to reset the heaps was to reboot.

You would have to compare your recollection, to concepts like this.
The tickless scheduler, which uses timers for keeping track of things.
Linux has an option like this too (it cost them a lot of changes to
their source code). The idea, is to no longer rely on steady clock tick
interrupts, to drive the scheduler. When the system is sleeping between
activities, it means the processor can go into a deeper C-state (and
be awakened by an aperiodic timer, rather than by a clock tick).

http://www.sandia.gov/tiros/

The problem, is finding good articles, that compare the characteristics,
and provide diagrams of how scheduler slots can be used.

A program can still "cooperate" with the OS. For example, if I write some
code, and do sleep() or issue a HLT instruction, that passes control from
my program back to the OS. Then, depending on how the scheduler works, either
the end of the scheduling period can be used for something else, or not...

It's possible your recollection is for preemption where scheduling slots
aren't being entirely used. And that would lead to slightly more waste.

Back in the day, another issue was the complexity of the scheduler. It
was possible to design "more fair" schedulers, or "more efficient" schedulers,
but you couldn't have both. The idea, was to keep the number of CPU cycles
used by the scheduler itself, to say, 3 to 5%, so the scheduler wouldn't be
wasting your available processor. I can't say now, how those metrics compare,
and how much waste there is on a modern processor, due to the scheduler itself.
A "strict priority" scheduler, versus a "round robin" scheduler, might have
had implications for meeting those kinds of performance targets. I think our
OS used "strict priority", and that was the designer justification of it
(slightly lower overhead, with a less sophisticated scheduler). The processors
at the time, might have been around 25MHz or so.

Paul

 

[Yes]

Thanks everyone.

As I mentioned, surfing the internet is slow right from the beginning
of reformat and install, and it just gets worse.

When I said "Ctrl+Alt+Del" doesn't work I was thinking in terms of
turning my PC off after it freezes up. (But obviously in XP that is
not what it is supposed to do). Nevertheless, it does usually brings
up the Task Manager, but that may freeze also, necessitating an
unplug.

In the "Task Manager" Firefox appears to use the most memory at over
235k. (Perhaps because I have a lot of tabs open at the moment). I
have 31 items under "Processes" and when I refresh a window it will
shoot up to 100% under the "Performance" tab.

Which version of FF? Keep in mind FF has always been a memory glutton
and in old versions people kept referring to an unidentified memory
leak that the user would have to shut down FF to reset memory, IIRC.

BTW. The system has Windows XP SP2.(I no longer bother with Windows
updates). I even uninstalled Google Chrome because of scripts which
kept trying to run even though I was only using Firefox.

I run WinXP, SP3 and recommend updating to it from SP2. I've been
using SP3 for so long that I can't tell you the exact things I saw
different between the two, but I use FF 10 with a number of add-ons for
my browsing and don't have the same issues you mention. FF still uses
more memory than seems reasonable to me, IMO.

As for hard drives, I've used different ones after each install, and
always have two in the system. A 20G and a 40G. (Not including the
500G USB drive). I'm only using about 1/3 of the space on my 40Gig "C"
drive.
http://i290.photobucket.com/albums/ll257/Statenislander/Computer/SequoiaView.jpg

I'll ditch AVG and go with either Avast! again or Avira.

A pointer to how to turn off processes would also be appreciated. I
wish there was a way to avoid all scripts which are a main culprit in
freezing up my system. I only get individual options to stop the
script from running until the next one pops up.

There are several options, the simplest using task manager. Under the
processes tab, task manager allows you to view a number of variables.
At the bottom is a clickbox that allows you to terminate a process -
Not really a good idea unless you know what the process does. You can
use the process name in google to discover more about what the process
does. I've found it useful to also pay attention to how many
processes are running - task manager tells you. Over time, you'll get
a feel for that. On my pc, it's usually around 32 whereas on my laptop
it's around 40. Those number are specific to my eqpt and will probably
differ from yours.

Alternatively, take a look at Microsoft's Sysinternals page. It has a
number of very good programs to help analyze and manage your O/S. I
don't have the URL handy.

Obviously this system cannot handle the internet efficiently anymore.
http://i290.photobucket.com/albums/ll257/Statenislander/Computer/BelarcAdvisor.jpg

I'm going to have to ditch this Dell XPS-Z Pentium III and build up
the hardware I already have, which include either a Dell 8400, HP
7955, or an ASUS A7V8X motherboard I can put in a case I have. All
have have a minimum of 2.8Ghz processors. I'll just have to make sure
the memory is maxed out or at least over 1G on whatever I system use,
and of course use a graphics card.

Your choice, but from the info you've given, it seems like upping your
memory to 1Gb might give it new legs. You'd have to balance that
against the cost of the replacement and upgrade of your eqpt. It is
always nice to have newer eqpt, but I'm still not sure the case has
been made that your computer cannot handle the net anymore.

I'd start with some basic things. One, upgrade to SP3. Two, check
your pagefile size settings. Three, doublecheck your FF settings
(because your post suggests that that is your default browser).
Perhaps test out how it feels if FF blocks automatic image downloads.
AFAIK, you have to have IE 8 installed for SP3, but you don't have to
use it. Use something such as Microsoft Sysinternals to look at what
starts up when you boot up your pc. Note that Sysinternals may require
a bit of learning curve; the various programs give a lot of info. I'm
by no means knowledgeable about everything it does, so I just focus on
the start up section. Four, use something like CCleaner to delete
trash off your hard drives and to scan for registry issues. Five,
doublecheck that your trash bin is empty. I for one keep forgetting
that moving something to trash bin does not delete the file; that has
to be done separately.

Those are free things that require spending time, not money. If you're
still not satisfied, then choose if you want to spend your money either
to buy more ram as a last ditch effort or to buy eqpt to replace/update
your pc.

 

[BillW50]

In

Searcher7 wrote:
I run WinXP, SP3 and recommend updating to it from SP2...

I'd start with some basic things. One, upgrade to SP3...

You know I have over a dozen XP machines here with almost all of them
have been updated to SP3 years ago. But if I had to do it all over
again, I don't think I would.

As most already know, making any changes to an OS carries a risk of
causing stability problems. And since the day Vista was released,
Microsoft has not shown any evidence that they care about any stability
problems that any security patch causes with XP. As when SP3 broke
compacting with OE6, Microsoft just didn't care and never bothered to
fix it. That is only one example, but you get the idea.

As far as security updates protecting you from infections, I too
believed this was true too. But over the years I started to have doubts
about this since I wasn't seeing any real evidence. I've been running
Windows since '93 and I never had an infection on any of my computers
yet. And on some of my computers I have stopped security updates for
years now and still I never had an infection. There must be a reason for
this?

 

[glee]

In snip
You know I have over a dozen XP machines here with almost all of them
have been updated to SP3 years ago. But if I had to do it all over
again, I don't think I would.

As most already know, making any changes to an OS carries a risk of
causing stability problems. And since the day Vista was released,
Microsoft has not shown any evidence that they care about any
stability problems that any security patch causes with XP. As when SP3
broke compacting with OE6, Microsoft just didn't care and never
bothered to fix it. That is only one example, but you get the idea.

SP3 didn't break compacting in OE. There is a Registry counter that
gives you a prompt to compact after 100 closings of OE. There were a
few programs that interacted with the counter after SP3 was installed,
causing the counter to increase more quickly than just with OE closings.
I've not seen the issue on any systems I have worked on since SP3 was
released, others have. Apparently some Nero plug-ins could increase the
count but I never saw it and I have been using Nero for years, the
Mailwasher program interfered, and supposedly installing Windows Live
Mail also affected the counter. The counter can be manually reset
through a registry edit, and Tom Koch made a tool that manually
compacted on demand and also reset the counter at the same time. None
of these are necessary, if you don't use Mailwasher or install WLM.

As far as security updates protecting you from infections, I too
believed this was true too. But over the years I started to have
doubts about this since I wasn't seeing any real evidence. I've been
running Windows since '93 and I never had an infection on any of my
computers yet. And on some of my computers I have stopped security
updates for years now and still I never had an infection. There must
be a reason for this?

Famous last words: I don't install security updates and I never had an
infection. But others have, even with updates installed, because so
many infections are due to social engineering. If you are smarter than
the average bear and practice Safe Hex, you are unlikely to get an
infection. Most users need the security updates. By not installing
them, you leave yourself open to getting pwned due to a security hole,
and you and your AV will never know.

 

[BillW50]

In

SP3 didn't break compacting in OE. There is a Registry counter that
gives you a prompt to compact after 100 closings of OE. There were a
few programs that interacted with the counter after SP3 was installed,
causing the counter to increase more quickly than just with OE
closings. I've not seen the issue on any systems I have worked on
since SP3 was released, others have. Apparently some Nero plug-ins
could increase the count but I never saw it and I have been using
Nero for years, the Mailwasher program interfered, and supposedly
installing Windows Live Mail also affected the counter. The counter
can be manually reset through a registry edit, and Tom Koch made a
tool that manually compacted on demand and also reset the counter at
the same time. None of these are necessary, if you don't use
Mailwasher or install WLM.

Oh boy! Yes SP3 did break compacting with OE6. As the newsgroups lit up
about this problem when SP3 was first released. And Bruce Hagen often
talked about it. I didn't see it at first, but I did finally run into it
(actually years after having SP3 installed). What I had seen happen is
it actually compacts just fine at first and the very last thing it tries
to update is folders.dbx. But it can't and an error message states it is
in use by another application. Here read this:

http://www.outlookforums.com/threads/45787-cannot-compact-files/

I don't actually follow Bruce's advice though. As all of the years I've
used OE, I almost never compact and I personally haven't ran into a
single problem not doing so. But I don't doubt for a second that others
can and do have problems. Oh yeah, my fix is to tell OE to go offline
and close OE down. Then reopen it and OE won't be doing anything but
just sit there. Now compact and don't do anything else with OE until it
is done. And that always worked for me so far.

And yes I know all about the counter and how it works and all. And I
might be wrong here, but I thought the counter thing was put in there by
SP2 and not SP3.

Famous last words: I don't install security updates and I never had an
infection. But others have, even with updates installed, because so
many infections are due to social engineering. If you are smarter
than the average bear and practice Safe Hex, you are unlikely to get
an infection. Most users need the security updates. By not
installing them, you leave yourself open to getting pwned due to a
security hole, and you and your AV will never know.

I understand all of this. And I base my opinion on testing on about 6
computers for over 5 years. And if not updating someday becomes fool
hardy, no problem. I have tons of backups to fix that problem really
quickly. But others should be doing this too if they want to stop
updating anyway.

And my beef about updates is that I am not troubled by malware, but
updates breaking something I am constantly fighting over. I have over 20
computers here and it doesn't take long before an update to screw up at
least one of them. So maybe you now know where I am coming from.

 

[glee]

In

Oh boy! Yes SP3 did break compacting with OE6. As the newsgroups lit
up about this problem when SP3 was first released. And Bruce Hagen
often talked about it. I didn't see it at first, but I did finally run
into it (actually years after having SP3 installed). What I had seen
happen is it actually compacts just fine at first and the very last
thing it tries to update is folders.dbx. But it can't and an error
message states it is in use by another application. Here read this:

http://www.outlookforums.com/threads/45787-cannot-compact-files/

I don't actually follow Bruce's advice though. As all of the years
I've used OE, I almost never compact and I personally haven't ran into
a single problem not doing so. But I don't doubt for a second that
others can and do have problems. Oh yeah, my fix is to tell OE to go
offline and close OE down. Then reopen it and OE won't be doing
anything but just sit there. Now compact and don't do anything else
with OE until it is done. And that always worked for me so far.

And yes I know all about the counter and how it works and all. And I
might be wrong here, but I thought the counter thing was put in there
by SP2 and not SP3.


I understand all of this. And I base my opinion on testing on about 6
computers for over 5 years. And if not updating someday becomes fool
hardy, no problem. I have tons of backups to fix that problem really
quickly. But others should be doing this too if they want to stop
updating anyway.

And my beef about updates is that I am not troubled by malware, but
updates breaking something I am constantly fighting over. I have over
20 computers here and it doesn't take long before an update to screw
up at least one of them. So maybe you now know where I am coming from.

Regarding OE compacting, we are talking about the same "issue"...
compacting was not "broken" however. Background compacting was the
problem, but compacting when closing OE or manually compacting while
offline worked fine.

I've never had a critical security update mess up a computer, unless the
computer was infected with a root kit or trojan. I'm not talking about
half a dozen computers, I'm talking about the neighborhood of 50 to 75.
I never updated while running other programs, and always disable the AV
during update installation. Most of my clients did not follow that
procedure though, and they still never had update problems. When people
tell these tales of critical updates killing their system, it has always
been due to other issues in every case I have seen. I'm curious....
when was the last time you checked you system for root kits and trojans
with Windows NOT loaded.... booting from a Linux-based AV rescue CD?

 

[Mark]

[-snip-]

As I mentioned, I have a laptop with Win7 and 2GB RAM using a Sempron
single core and no problems at all. It helps to keep the background
programs and processes under control. I've seen machines with a
ridiculous number of background processes running for no good reason.

Most PCs ship with massive amounts of bloatware preinstalled. The
first job is to remove it all. I recently bought a W7 laptop and,
when no applications were running, it was using over 3G of memory! We
got rid of loads of crap and it's much better but still uses a lot
more than XP did.

 

[Mark]

If you don't know the brand of your hard drive, you can use Hitachi
Drive Fitness Test (DFT) on almost any brand drive.
http://www.hitachigst.com/support/downloads/#DFT

Does this work with RAID controllers?