stop over complicating things

Ian Betts wrote:
>
>
> "Charlie Tame" <charlie@tames.net> wrote in message
> news:#u9R#$c1HHA.2752@TK2MSFTNGP06.phx.gbl...
>> Ian Betts wrote:
>>>
>>>
>>> "Charlie Tame" <charlie@tames.net> wrote in message
>>> news:uPTXrWY1HHA.1344@TK2MSFTNGP04.phx.gbl...
>>>> cquirke (MVP Windows shell/user) wrote:
>>>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:
>>>>>
>>>>
>>>>> Vista may run software faster than XP, in the context of mid-Vista
>>>>> hardware (e.g. multi-core processors, 2-4G RAM, etc.).
>>>>
>>>> Nope, slower, quite a bit slower. 3 different machines - all 4GB and
>>>> dual processor, all slower than XP. Sorry. We have gone from saying
>>>> "It's because your machine is out of date" to saying "It's because
>>>> your current model machine is out of date".
>>>>
>>> Two machines, exactly the same spec , side by side, one XP the other
>>> Vista. No appreciable difference although the Vista is a tad faster
>>> in fact. It is not about the processor, it is about the full spec on
>>> the comp.
>>>
>>>
>>>
>>
>>
>> Yep, so any user has a problem it's THEIR fault. As above it's either
>> the machine or something the user has done, despite the fact that you
>> can never actually PROVE where the alleged problem lies.
>>
>> Now, instead of simply implying there is something wrong with the
>> machines of just about every person who posts here, despite the fact
>> that many are brand new machines "Designed for Vista" etc and
>> reasonably clean new installs why not try to actually figure out
>> what's going on.
>>
>> Quite honestly I am not surprised that the Linux community give you a
>> hard time in this group, Vista has gotten off to a less than perfect
>> start and the only response most here give is that it's hardware,
>> drivers or the user... for example the recent post that you obviously
>> didn't even read.
>>
>> I have older motherboards that run well with Vista and XP but always
>> XP is faster. I have brand new motherboards that run well with Vista,
>> but XP is always faster. My conclusion in your case is that either the
>> XP install is older and maybe has problems, or maybe it is simply
>> running something else that you are not mentioning or perhaps are
>> unaware of. It also depends somewhat on what software one is trying to
>> run. For pure OS stuff Vista is slower, for the likes of games it is
>> seriously slower, and for file copying etc that involves interaction
>> between machines it appears to be a real loser.
>>
>> Simply making broad generalizations about machine spec helps nobody,
>> especially Microsoft. Not everybody who comes here asking questions is
>> completely devoid of knowledge.
>>
> And you are now the expert. Just where have I ever been dismissive of a
> question. I did not say it was what you had done but that you were wrong
> in your claim, oft repeated. I do not answer those questions I cannot
> give an answer to, nor do I tell lies. The two computer have exactly the
> same software except, one has the last, I.E. XP Pro 64bit 2006 version
> and the second has the Latest Premium Vista. Try to think before you
> criticize those who help as opposed to those who just condemn like you,
>


You have that exactly the wrong way around Ian. First I have not "Oft
repeated" anything except to confirm suggestions made by others, the
vast majority of others who post here with problems (Which does not
imply the vast majority of users) that relate to Vista's slow response
times, dreadful frame rates and apparent trips off to never never land
whilst showing the "Busy" icon.

So, you are calling me a liar saying I am wrong in my claim, yet I have
the evidence that VISTA IS SLOWER than XP on 3 machines. You expect me
to lie about what I see in order to hide facts you prefer not to see?

I am not in the habit of criticizing those who help, nor do I set out to
condemn. There's nothing I can do about the facts, whether you like them
or not, but frankly it is NOT helping users when they are simply told
"It's not the OS, it is you or your machine or something we can't answer"

Perhaps YOU should reflect on the quality of the "Help" given before
accusing others of criticizing it.

Last evening there was a real live MS employee here, he noted a number
of problems mentioned by users and left them with a positive feeling
that at least someone had heard the complaint and it would get some
attention. Needless to say he can't do it all on his own and will
probably tire quickly if that's what happens. What was clear is that
nobody expects an OS to be perfect, but problems need to be discussed
and chased down. Quite where blaming the users / hardware / drivers fits
into this I do not know.

MS need to know what irritates customers, it is not serving them or
anyone else to pretend, or to try and harrass others into pretending the
problems don't exist.

On Thu, 02 Aug 2007 23:20:41 -0500, Charlie Tame wrote:
>cquirke (MVP Windows shell/user) wrote:
>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" wrote:

>> Vista may run software faster than XP, in the context of mid-Vista
>> hardware (e.g. multi-core processors, 2-4G RAM, etc.).

>Nope, slower, quite a bit slower.

Duno how you're testing on "mid-Vista hardware" in this, the first
year of Vista. We're talking mid-2009 hardware, here (or later).

>We have gone from saying "It's because your machine is
>out of date" to saying "It's because your current model
>machine is out of date".

This is an OS designed for the next 5 - 10 years, so a budget 2007 PC
is, in that context, the lamest PC Vista is likely to see.



>------------ ----- ---- --- -- - - - -
The most accurate diagnostic instrument
in medicine is the Retrospectoscope
>------------ ----- ---- --- -- - - - -

cquirke (MVP Windows shell/user) wrote:
> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:
>
>> It's not just about games. Overall performance in applications and
>> such is generally slower and poorer in Vista. Again, I just see no
>> excuse for this, considering everything Vista/Longhorn promised to
>> bring.
>
> A new OS is about scalability, not "speed".

Yes, it's about scalability, but that's not JUST what it's about. Vista
(as Longhorn) originally promised speed as well, and so far hasn't come
close to delivering. Hell, it hasn't really delivered on the scalability
either.

> Do 1998 programs run faster on a 1998 PC with 64M RAM running Windows
> 98, or Windows 2000?

In my experience most ran better on 2000, unless you're talking about
Win 2000 prereleases.

> Shouldn't Windows 2000 automatically be faster, as it's all-32-bit?
>
> Now try the same test on a PC with 1G RAM.
>
> Now try the same test with a 2008 program that requires 512M RAM to
> operate, as well as > 137G HD space. How good is Win98 now?


Thats part of the problem as well; too many programs are far over weight
in regards to the resources they require. Many programs (Office,
Acrobat, etc) don't really do much more than much older versions did for
the most part, yet they have grown to consume around 10 fold the
resources. What reason is there for this? (I suspect that newer dev kits
being partly to blamed; the .Net frame works, for example, are seriously
bloated. I coudl write two identicle prgorams in VC++ 6 or lower and in
VS 2003 or 2005 and the former would be infinately smaller. Theres just
no reason for so much extra fat.)

> Vista may run software faster than XP, in the context of mid-Vista
> hardware (e.g. multi-core processors, 2-4G RAM, etc.). Vista's not
> built for 2007 and software of that vintage; it has to support several
> years of future development, which XP is unlikely to do as well.

I disagree. The original promises of then Longhorn, now Vista, were that
it weould be faster than XP and previous versions, would be able ot run
almost anything. Instead we have a lardass of an OS laden with DRM and a
mountain of compatibility issues, making it a far cry from it's
promises.

> Windows XP does not run DOS games as well as Win98. In 2007, do you
> care? In 2015, will you care whether Vista runs XP-era software as
> well as XP would have done?

While many say this is as muhc the fault of soundcard manucatures (sound
being the biggest problem running old DOS games under XP, and video also
being a factor for games that require extended VESA support.) IMHO,
Microsoft dropped the ball on that one. Then again, many classic dos
games (Doom, Warcraft, etc) have had Windows versions for some time, and
simpler games could be run under "DosBox" as well, which gives you far
more features the stock cmd.exe ever could.

Still, I do wish Microsoft had given better attention to legacy DOS
emulation in XP. But Vista is far worse, as it makes many perfectly good
Win32 apps unusable, which IMHO is just unexcuable for a "next-gen" 32
bit Windows. (And 64 bit is even more a mess in this regard on Vista as
well as XP.)



In the end, I find completely inexcusable that an OS as new as Vista, on
3+ gz dual core processor to be so slow. I cna compare it to an old
pentium 333 mhz 2 laptop I have laying running Win 2000 that can run
most applications about as well as a typical Vista machine I've used in
any shop or when working on customer's machines. Ok, maybe thats a
slight exaggeration, but still, I expect so much more speed and power
with todays hardware. FSB speeds are slowly catchign up, yet OS's like
Vista just don't give you the speed and laglessness you'd expect. Going
from a 386 to a 486 was quite a boost... so was a 486 to a P166 or 200,
and thne to a p2 400 or p3 550. Why then is performance so damn lacking
between a 550 p3 and 3gz Dual Core running Vista. XP on the same 3 gz
Dual Core gives a bit more speed, although I'd agree it should be better
too. Bust with Vista theres simply NO EXCUSE.

-saran

Ian Betts wrote:
> "Charlie Tame" <charlie@tames.net> wrote in message
> news:#u9R#$c1HHA.2752@TK2MSFTNGP06.phx.gbl...
>> Ian Betts wrote:
>>>
>>>
>>> "Charlie Tame" <charlie@tames.net> wrote in message
>>> news:uPTXrWY1HHA.1344@TK2MSFTNGP04.phx.gbl...
>>>> cquirke (MVP Windows shell/user) wrote:
>>>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:
>>>>>
>>>>
>>>>> Vista may run software faster than XP, in the context of mid-Vista
>>>>> hardware (e.g. multi-core processors, 2-4G RAM, etc.).
>>>>
>>>> Nope, slower, quite a bit slower. 3 different machines - all 4GB
>>>> and dual processor, all slower than XP. Sorry. We have gone from
>>>> saying "It's because your machine is out of date" to saying "It's
>>>> because your current model machine is out of date".
>>>>
>>> Two machines, exactly the same spec , side by side, one XP the other
>>> Vista. No appreciable difference although the Vista is a tad faster
>>> in fact. It is not about the processor, it is about the full spec
>>> on the comp.
>>>
>>>
>>>
>>
>>
>> Yep, so any user has a problem it's THEIR fault. As above it's
>> either the machine or something the user has done, despite the fact
>> that you can never actually PROVE where the alleged problem lies.
>>
>> Now, instead of simply implying there is something wrong with the
>> machines of just about every person who posts here, despite the fact
>> that many are brand new machines "Designed for Vista" etc and
>> reasonably clean new installs why not try to actually figure out
>> what's going on. Quite honestly I am not surprised that the Linux
>> community give you
>> a hard time in this group, Vista has gotten off to a less than
>> perfect start and the only response most here give is that it's
>> hardware, drivers or the user... for example the recent post that
>> you obviously didn't even read. I have older motherboards that run
>> well with Vista and XP but always
>> XP is faster. I have brand new motherboards that run well with
>> Vista, but XP is always faster. My conclusion in your case is that
>> either the XP install is older and maybe has problems, or maybe it
>> is simply running something else that you are not mentioning or
>> perhaps are unaware of. It also depends somewhat on what software
>> one is trying to run. For pure OS stuff Vista is slower, for the
>> likes of games it is seriously slower, and for file copying etc that
>> involves interaction between machines it appears to be a real loser.
>>
>> Simply making broad generalizations about machine spec helps nobody,
>> especially Microsoft. Not everybody who comes here asking questions
>> is completely devoid of knowledge.
>>
> And you are now the expert. Just where have I ever been dismissive of
> a question. I did not say it was what you had done but that you were
> wrong in your claim, oft repeated. I do not answer those questions I
> cannot give an answer to, nor do I tell lies. The two computer have
> exactly the same software except, one has the last, I.E. XP Pro 64bit
> 2006 version and the second has the Latest Premium Vista. Try to
> think before you criticize those who help as opposed to those who
> just condemn like you,

You can't compare a 64 bit XP to a 32 bit Vista. 64 Bit XP/Vista will
always be slower with 32 bit programs because of the emulation layer
(since addressing is natively 64, not 32 bit.)

When you compare 32 bit XP to 32 bit Vista, assuming both installs are
healthy and aren't running something that can arbitrarily slow one down,
XP will virtually always be faster.


-saran

Frank wrote:
> Alias wrote:
>
>> Ian wrote:
>>
>>> I think you may be in for a shock. XP will not be supported for more
>>> than five years and more likely two by MS and many processors will
>>> out distance lesser bistros. Printer markers will soon stop supply
>>> update drivers because they make no profit from printers and rely
>>> on ink sales. Video cards will have to meet bigger demands and the
>>> older ones will not be supported and that is only half of it.
>>
>>
>> Great reasons for going for Open Source.
>>
> Best reason I know of to make the move to the very best os there
> is...Vista! Frank

Well I hope you enjoy your DRM and slower computing experience.
Meanwhile everyone who actually know what they are doing will be running
an OS (be it XP, 2000, 2003 or something else) and running circles
around you for years to come.

-saran

On Thu, 9 Aug 2007 11:34:07 -0700, "Saran" <none@nospam> wrote:
>cquirke (MVP Windows shell/user) wrote:
>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:

>> A new OS is about scalability, not "speed".

>Yes, it's about scalability, but that's not JUST what it's about. Vista
>(as Longhorn) originally promised speed as well, and so far hasn't come
>close to delivering. Hell, it hasn't really delivered on the scalability
>either.

Now that you mention it, Windows 4 was going to deliver speed when it
was first announced, too. We expected it, as on paper, 32-bit flat
memory addressing and no separate DOS layer sounded way more
efficient... but by the time Windows 4 finally blinked into the light
of day as NT 3.1, it was a resource-heavy slug that needed four times
the RAM just to break even in speed.

I guess when Vista was still young enough to be called Longhorn, such
claims may have been made, too. I dunno if I don't remember them
because I wasn't interested in it at that time, or whether I heard
them and rejected such claims before they reached my memory ;-)

In practice, Vista does have a lot of speed-up technologyy, but this
serves only to soften the body-blows from its huge overheads, much as
Windows 3.1's aggressive write caching (with all the file system
corruption that implies) merely help it be slightly less grossly
slower than DOS as an application platform.

>> Do 1998 programs run faster on a 1998 PC with 64M RAM running Windows
>> 98, or Windows 2000?

>In my experience most ran better on 2000, unless you're talking about
>Win 2000 prereleases.

No; thinking Win2000 RTM and SP* - but, in only 64M RAM. I might
agree "better" in stability and uptime terms, though.

>> Now try the same test on a PC with 1G RAM.
>>
>> Now try the same test with a 2008 program that requires 512M RAM to
>> operate, as well as > 137G HD space. How good is Win98 now?

>Thats part of the problem as well; too many programs are far over weight
>in regards to the resources they require.

Hardware (made in China) is cheap, whereas software (made at least
partly in USA) is not - and this is more than a labor wage thingl;
there's a genuine complexity factor, too.

So the trend is to throw hardware at speeding up software, rather than
boost software development efforts to better fit smaller hardware.

This may change, given "green" concerns, and/or a move away from
formal PCs to smaller devices as users lose interest in bigger sware.
It's becoming less and less clear that more software makes things
better; Vista almost collapses under the weight of all the "computing
about computing" that it does, and at least some user feedback
suggests the benefits of this are not outweighing the costs.

>Many programs (Office, Acrobat, etc) don't really do much more
>than much older versions did for the most part, yet they have
>grown to consume around 10 fold the resources.

Yup. Software vendors make software, and thus money.
They ain't gonna get rich making less software ;-)

>What reason is there for this? (I suspect that newer dev kits
>being partly to blamed; the .Net frame works, for example, are seriously
>bloated. I coudl write two identicle prgorams in VC++ 6 or lower and in
>VS 2003 or 2005 and the former would be infinately smaller. Theres just
>no reason for so much extra fat.)

Actually, there is, if this "fat" formalizes the layers and surfaces
between software components so as to reduce bugs and exploits.

"Never code anything bigger than your own head"

Once a project is too big to fit in a single dev's head, the era of
"artistic programming" (as in "I can't understand a word of the source
code, but wow, the object code kicks butt!") is over.

Attention turns instead to formalizing the way different devs see and
(re-)use each other's work, in an effort to kill bugs. The process of
killing these bugs is rarely over by the time the product ships
(currently, we expect better from hardware development, which is why
we are more confident in large hardware carrying the load).

If the kill-the-bugs phase is never over, the optimize-for-speed
process never even begins. In fact. there's so much fear of
regression bugs (new bugs introduced by "improving" the code) that it
may be rare to re-cast working code just to improve efficiency.

>> Vista may run software faster than XP, in the context of mid-Vista
>> hardware (e.g. multi-core processors, 2-4G RAM, etc.). Vista's not
>> built for 2007 and software of that vintage; it has to support several
>> years of future development, which XP is unlikely to do as well.

>I disagree. The original promises of then Longhorn, now Vista, were that
>it weould be faster than XP and previous versions, would be able ot run
>almost anything. Instead we have a lardass of an OS laden with DRM and a
>mountain of compatibility issues, making it a far cry from it's promises.

I don't think I ever took "it will be faster" claims seriously, if
indeed I ever heard them. However, I do expect to see 2009-era
software that works better on Vista than XP, and 2010 software running
faster on Vista64 than any XP, and 2112 software that requires Vista
and won't run on XP at all.

The time frames may change, driven largely by external events; malware
evolution and to what extend DRM-riddled media becomes pervasive.

>> Windows XP does not run DOS games as well as Win98. In 2007, do you
>> care? In 2015, will you care whether Vista runs XP-era software as
>> well as XP would have done?

>While many say this is as muhc the fault of soundcard manucatures (sound
>being the biggest problem running old DOS games under XP, and video also
>being a factor for games that require extended VESA support.) IMHO,
>Microsoft dropped the ball on that one.

DOS-era hardware is so different, compared to modern stuff - and the
speed difference is so great - that the best approach to running the
stuff may be to emulate it, as if it was an alien platform - much as
one might emulate a ZX Spectrum on a 200MHz Pentium.

There's so much spare speed and capacity that one can not only "keep
up" with the DOS app's needs, one can go further and deliberately
emulate accurate per-instruction timings, etc.

>Then again, many classic dos games (Doom, Warcraft, etc) have
>had Windows versions for some time

Ain't the same, sometimes. Wolfenstein isn't the same without the
flat ceilings, lurid primary colors, chintzy FM music synthesis and
gritty 8-bit sound samples ;-)

>simpler games could be run under "DosBox" as well, which gives you far
>more features the stock cmd.exe ever could.

Yep, that's what I had in mind. Eventually, sub-PC devices may get
fast enough to port the lot over to a PC emulator on a cell phone.

>In the end, I find completely inexcusable that an OS as new as Vista, on
>3+ gz dual core processor to be so slow.

It's not that far outside my expectations, especially if you're
looking at 5-year OS cycles instead of 3-year NT or 2-year Win9x.

XP on 2002 PCs was pretty heavy, too, as was Win3.1 on the 2M or 4M
386 systems of their day. I still have clients using 2002-era XP
boxen with 128M RAM, and once they take my advice to do the RAM
shuffle (PC1 = new 512M, PC2 = 2 x 128M) they like++

XP came out as we moved from PIII to P4, Win98 from Pentium to PII,
Win95 from 486DXn to Pentium, and Win3.1 from 386 to 486. In each
case, you were caught between the exhausted tail of an old hardware
platform, and the feeble first of the next generation.

You will look back at today's sub-2GHz Core 2 Duos with the same sort
of semi-disgust as one might have considered a Pentium-66 or 486SX-25;
"at least it's a Core 2 / Pentium / 486" vs. "hell, it's so lame, it
may as well have been a P4 / 486DX4 / 386DX-40".

At least today's Core 2 platform has legs for the future - it's going
to clock and core scale way better than P4 could ever have done.

The same may well apply to Vista... do you remember the first
iterations of other significant "long legs for the future"
technologies? The years of "Useless Serial Bus", the costs and buggy
LAN cards of PCI, and all the grumbles about Plug-n-Play?

But we'll only know when we get there. In 2007, we can only guess.

>I cna compare it to an old pentium 333 mhz 2 laptop I have
>laying running Win 2000 that can run most applications about
>as well as a typical Vista machine I've used in any shop

Yup. I had a Pentium-200 running Win95 "build 950" in 8M RAM, and the
old thing rocked surprisingly well. To me, that was the breakthrough
OS, just as the original 386 was the breakthrough processor.

>Going from a 386 to a 486 was quite a boost...

Not really, no - in fact, initially the non-Intel wolves would rip the
flesh off Intel's initial overpriced offerings. AMD 386DX-40 or Intel
486SX-25? AMD 586DX4-133 or Pentium-75?

But once the 486 discovered core multipliers to become the DX4, and
Pentium routinely enjoyed 66MHz base speed and lower operating
voltages, that's when things started to rock. Expect n-core
processors (and an OS that knows what to do with them), etc.

>so was a 486 to a P166 or 200

That's exactly what I mean. When the 486 hadn't discovered core speed
multiplication or sub-5v power yet, the Pentium SUCKED (do you
remember the appalling P60 and P66?).

When the Pentium Pro tried dedicated 32-bit design for the first time,
it was effective, but niched due to costly manufacture and poor 16-bit
speed. In fact, it took a while before Intel shrunk the silicon
enough to fit the whole processor core in a single die, and it was
only then that the PII/III really lived up to its potential.

So we've seen processor generation false starts before, even with
fairly sub-revolutionary generations (think PIII Slot One vs. S370,
and S426 ? vs. S478 P4). We may yet see the same thing with the Core
2 generation, though this time it looks good from Day 1.



>------------ ----- ---- --- -- - - - -
The most accurate diagnostic instrument
in medicine is the Retrospectoscope
>------------ ----- ---- --- -- - - - -

cquirke (MVP Windows shell/user) wrote:
> On Thu, 9 Aug 2007 11:34:07 -0700, "Saran" <none@nospam> wrote:
>> cquirke (MVP Windows shell/user) wrote:
>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:
>
>>> A new OS is about scalability, not "speed".
>
>> Yes, it's about scalability, but that's not JUST what it's about.
>> Vista (as Longhorn) originally promised speed as well, and so far
>> hasn't come close to delivering. Hell, it hasn't really delivered on
>> the scalability either.
>
> Now that you mention it, Windows 4 was going to deliver speed when it
> was first announced, too. We expected it, as on paper, 32-bit flat
> memory addressing and no separate DOS layer sounded way more
> efficient... but by the time Windows 4 finally blinked into the light
> of day as NT 3.1, it was a resource-heavy slug that needed four times
> the RAM just to break even in speed.

This is true, but it seems this also shows Microsoft hasn't learned much
from those days. Also, NT was never marked for home use like Vista is
for the most part is. It was for business use and business setups
typically use more resources depending on the type of work than home
users.

> I guess when Vista was still young enough to be called Longhorn, such
> claims may have been made, too. I dunno if I don't remember them
> because I wasn't interested in it at that time, or whether I heard
> them and rejected such claims before they reached my memory ;-)

Again, it appears Microsoft has learned little in the grand scheme of
things, other than how to make more money with inferior but flashy
goods. Its not so different from how some auto makers routinely cut
corners in production and generally crank out cars that are inferior
than previous models.

> In practice, Vista does have a lot of speed-up technologyy, but this
> serves only to soften the body-blows from its huge overheads, much as
> Windows 3.1's aggressive write caching (with all the file system
> corruption that implies) merely help it be slightly less grossly
> slower than DOS as an application platform.

Its the huge overhead in Vista and many new applications that's the
problem. Too much over head and too little new/gain. That sort of
equation doesn't work in general cost benefit analysis, so why on earth
should Vista be any different? Too much cost, in terms of money for the
OS itself plus the hardware needed just to run it adequately and time
spend dealing with all the issues as well as the extra overhead itself,
and too little benefit over previous versions.

>>> Do 1998 programs run faster on a 1998 PC with 64M RAM running
>>> Windows 98, or Windows 2000?
>
>> In my experience most ran better on 2000, unless you're talking about
>> Win 2000 prereleases.
>
> No; thinking Win2000 RTM and SP* - but, in only 64M RAM. I might
> agree "better" in stability and uptime terms, though.

It depends on what you're trying to run. Normal applications ran so much
better. The improved memory management over 9x was worth it alone for
me, and that also translated to far fewer reboots needed (since I left
9x years ago I hardly ever need to reboot any system I have.) If you
mean games, the initial versions of 2000 didn't fair too well but was
fixed later. I never saw any such problems in XP though, though that's
not to say there weren't any, just nowhere near the frequency of 2000's
initial release or Vista for that matter.

It seems to me that going from 2000 to XP, Microsoft really seemed to be
showing they learned a few things and the response was relatively
positive. Vista seems like they took a step back in many ways with all
the many issues. Why break so many things when they were working

>>> Now try the same test on a PC with 1G RAM.
>>>
>>> Now try the same test with a 2008 program that requires 512M RAM to
>>> operate, as well as > 137G HD space. How good is Win98 now?
>
>> Thats part of the problem as well; too many programs are far over
>> weight in regards to the resources they require.
>
> Hardware (made in China) is cheap, whereas software (made at least
> partly in USA) is not - and this is more than a labor wage thingl;
> there's a genuine complexity factor, too.

So why should one pay high prices when there isn't much new/benefit over
the previous versions? It's again the upgrading for the sake of
upgrading and it's just worse when they force you to upgrade (think AV
and Tax type software.)

> So the trend is to throw hardware at speeding up software, rather than
> boost software development efforts to better fit smaller hardware.

And you see nothing wrong with that?

> Yup. Software vendors make software, and thus money.
> They ain't gonna get rich making less software ;-)

Too many make the same software over and over and pitch it as something
new when it's essentially the same old thing. Some new version are even
worse, changing what made them so great. ACDSee used to be great, for
example. I'm tired of being told I need to upgrade or I'm gonna "lose
out" when a previous version does the job perfectly well. Why real
service does it to for you to install a new version that's eating more
resources and doing nothing real new. It just has a face lift, and maybe
a rearranged GUI, which, once again, is not necessarily a good thing.
This is also an issue in Vista: make changes for the sake of making
changes, changing things like labels that have been a particular way for
well over a decade.

>> What reason is there for this? (I suspect that newer dev kits
>> being partly to blamed; the .Net frame works, for example, are
>> seriously bloated. I coudl write two identicle prgorams in VC++ 6 or
>> lower and in VS 2003 or 2005 and the former would be infinately
>> smaller. Theres just no reason for so much extra fat.)
>
> Actually, there is, if this "fat" formalizes the layers and surfaces
> between software components so as to reduce bugs and exploits.

But it's introduced even more bugs and there already been news about new
exploits.

> "Never code anything bigger than your own head"
>
> Once a project is too big to fit in a single dev's head, the era of
> "artistic programming" (as in "I can't understand a word of the source
> code, but wow, the object code kicks butt!") is over.
>
> Attention turns instead to formalizing the way different devs see and
> (re-)use each other's work, in an effort to kill bugs. The process of
> killing these bugs is rarely over by the time the product ships
> (currently, we expect better from hardware development, which is why
> we are more confident in large hardware carrying the load).
>
> If the kill-the-bugs phase is never over, the optimize-for-speed
> process never even begins. In fact. there's so much fear of
> regression bugs (new bugs introduced by "improving" the code) that it
> may be rare to re-cast working code just to improve efficiency.

Well, I partly agree, but I see too many broken things in Vista for that
to really be the reality as it stands. What I see is (possibly with
every new release of VS and/or .Net Framework) things just get bigger
and fatter and slower for little gain between versions.

>>> Vista may run software faster than XP, in the context of mid-Vista
>>> hardware (e.g. multi-core processors, 2-4G RAM, etc.). Vista's not
>>> built for 2007 and software of that vintage; it has to support
>>> several years of future development, which XP is unlikely to do as
>>> well.
>
>> I disagree. The original promises of then Longhorn, now Vista, were
>> that it weould be faster than XP and previous versions, would be
>> able ot run almost anything. Instead we have a lardass of an OS
>> laden with DRM and a mountain of compatibility issues, making it a
>> far cry from it's promises.
>
> I don't think I ever took "it will be faster" claims seriously, if
> indeed I ever heard them. However, I do expect to see 2009-era
> software that works better on Vista than XP, and 2010 software running
> faster on Vista64 than any XP, and 2112 software that requires Vista
> and won't run on XP at all.

Maybe, maybe not. It's a lot of work to rework applications that
depended on 32 bit addressing. You can't necessarily just load up the
source on a 64 bit machine and hit compile. you have to take care with
data structures that used 32 bit types as recompiling doubles that
width, and thus takes up more space. Actually, it seems that this would
cause many programs to eat even more resources.

At some point you have to ask yourself something? Do you want faster,
more responsive software with a decently small footprint. Do we really
NEED all these new versions if the previous gets the job done better,
which a lot of the time seems to be the case these days. IMHO, I'd take
speed, efficiency and ease of use over eye candy any day.

> The time frames may change, driven largely by external events; malware
> evolution and to what extend DRM-riddled media becomes pervasive.

I think one could say all that DRM is actually fueling the piracy that
it want to stop. It seems to me torrent downloading of movies and TV
shows is at an all time high, and if that's so, I wouldn't be surprised.
I personally would rather have freedom than be shackled.

>> Then again, many classic dos games (Doom, Warcraft, etc) have
>> had Windows versions for some time
>
> Ain't the same, sometimes. Wolfenstein isn't the same without the
> flat ceilings, lurid primary colors, chintzy FM music synthesis and
> gritty 8-bit sound samples ;-)

Agreed. I enjoy the orignal everyone once in a blue moon via DosBox, as
it runs flawlessly there under XP at least.

>> simpler games could be run under "DosBox" as well, which gives you
>> far more features the stock cmd.exe ever could.
>
> Yep, that's what I had in mind. Eventually, sub-PC devices may get
> fast enough to port the lot over to a PC emulator on a cell phone.

I would like to see that

>> In the end, I find completely inexcusable that an OS as new as
>> Vista, on 3+ gz dual core processor to be so slow.
>
> It's not that far outside my expectations, especially if you're
> looking at 5-year OS cycles instead of 3-year NT or 2-year Win9x.

From what I understood, Longhorn goes back at least as far as 2001, so I
estimate a 6+ year overall development. Combined with the fact they
already had three rather well working OSes in 2000, XP, and 2003, yet
Vista does many things all that much worse. With Vista it seems to keep
coming back to upgrading for the sake of upgrading. 98 to 2000, 2000 to
XP, offered truly new and useful advancements. What does Vista really
offer that is really NEW ?

> XP on 2002 PCs was pretty heavy, too, as was Win3.1 on the 2M or 4M
> 386 systems of their day.

I never had any trouble running Win 3.1 on a 386/33 with 4M of ram

As for XP... it seems to run decently on a variety of systems. I run it
on a P2 laptop with 400M ram using SP2 and it's incredibly solid. Gaming
and high res videos aside, in general applications it's almost on par
with a well oiled p4 I also having running XP. At one time, as an
experiment, I even installed XP SP1 on P1 200 with 128M ram, and it ran
surprisingly well. Sluggish? Of course, but stable as hell, while the
initial startup was nothing to write home about, it still ran relatively
smooth and consistent, even if loading applications was rather slow. It
still worked and was surprisingly usable.

Can the same be said about Vista, in the relative sense? Hell, you need
the new high horse power just to make it run well. I'm still at a lose
as to why Vista on a decent 3+ GHz P4 or Dual Core almost feels like XP
on that P2 laptop? Again, this feels more like a step back than true
progress.

> XP came out as we moved from PIII to P4, Win98 from Pentium to PII,
> Win95 from 486DXn to Pentium, and Win3.1 from 386 to 486. In each
> case, you were caught between the exhausted tail of an old hardware
> platform, and the feeble first of the next generation.

386 to 486 to early Pentium was never so bad. There was some speed
difference (and more ram helped too.) Clock speeds seemed to jump around
30~ MHz at a time and gave decent gains. When then, with speeds in
excess of 100 fold (clock wise) from a 33 MHz 386 to a 3+ GHz
P4/DualCore/etc, performs so slow in comparison. What happened to all
the gain? 100 times faster in clock speed. 15 times that of a 200 MHz P1
or 6 times that of a 500 MHz P3. Yes, I know FSB speeds have been
struggling to catch up, but still, for such clock speed increases I
expect a lot more out of an OS that's supposed to be so advanced, in
that respect.

>> I cna compare it to an old pentium 333 mhz 2 laptop I have
>> laying running Win 2000 that can run most applications about
>> as well as a typical Vista machine I've used in any shop
>
> Yup. I had a Pentium-200 running Win95 "build 950" in 8M RAM, and the
> old thing rocked surprisingly well. To me, that was the breakthrough
> OS, just as the original 386 was the breakthrough processor.
>
>> Going from a 386 to a 486 was quite a boost...
>
> Not really, no - in fact, initially the non-Intel wolves would rip the
> flesh off Intel's initial overpriced offerings. AMD 386DX-40 or Intel
> 486SX-25? AMD 586DX4-133 or Pentium-75?
>
> But once the 486 discovered core multipliers to become the DX4, and
> Pentium routinely enjoyed 66MHz base speed and lower operating
> voltages, that's when things started to rock. Expect n-core
> processors (and an OS that knows what to do with them), etc.

The problem was that AMD wasn't so well known yet. Intel was the defacto
at the time, and yes the 486 DX2 66 w/ 8M ram (which is what I had back
in the day) kicked ass

>> so was a 486 to a P166 or 200
>
> That's exactly what I mean. When the 486 hadn't discovered core speed
> multiplication or sub-5v power yet, the Pentium SUCKED

Agreed.

> (do you remember the appalling P60 and P66?).

*Shudder* I try not to

> When the Pentium Pro tried dedicated 32-bit design for the first time,
> it was effective, but niched due to costly manufacture and poor 16-bit
> speed. In fact, it took a while before Intel shrunk the silicon
> enough to fit the whole processor core in a single die, and it was
> only then that the PII/III really lived up to its potential.
>
> So we've seen processor generation false starts before, even with
> fairly sub-revolutionary generations (think PIII Slot One vs. S370,
> and S426 ? vs. S478 P4). We may yet see the same thing with the Core
> 2 generation, though this time it looks good from Day 1.


Yes, but I really don't see how this really changes the facts about the
current state of the software industry cranking new resource hungry
version don't give much benefit over the version that came before.
Again, I tend to place Vista in this category based on the overhead,
changes for change sake, DRM, etc.

I for one don't like being treated as a sheep by vendors who expect me
to buy every new release just because it's new. If there's some real
benefit then it may be worth getting it and it's fairly priced, then it
makes sense. But as long as vendors (including Microsoft) release
over-priced, over-weight software with out giving some really benefit
that makes it worth it, then I'm not going to buy it. Simple as that.

-saran

cquirke (MVP Windows shell/user) wrote:
> On Thu, 02 Aug 2007 23:20:41 -0500, Charlie Tame wrote:
>> cquirke (MVP Windows shell/user) wrote:
>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" wrote:
>
>>> Vista may run software faster than XP, in the context of mid-Vista
>>> hardware (e.g. multi-core processors, 2-4G RAM, etc.).
>
>> Nope, slower, quite a bit slower.
>
> Duno how you're testing on "mid-Vista hardware" in this, the first
> year of Vista. We're talking mid-2009 hardware, here (or later).
>
>> We have gone from saying "It's because your machine is
>> out of date" to saying "It's because your current model
>> machine is out of date".
>
> This is an OS designed for the next 5 - 10 years, so a budget 2007 PC
> is, in that context, the lamest PC Vista is likely to see.



So you are telling me that an OS that runs quite a bit slower than XP on
several incarnations of the most recent hardware available to the public
is suddenly going to run faster as faster hardware becomes available...
isn't that rather a safe bet on your part?

On a 32 bit machine it is slower than XP.

On a 64 bit machine Vista 32 is slower than XP 32, Vista 64 is slower
than XP 64.

Now, W2000 was slower than 98 but we both know there are many good
reasons for that, and there are reasons why Vista is slower than XP but
I question the "Good" part.

Saran wrote:
> cquirke (MVP Windows shell/user) wrote:
>> On Thu, 9 Aug 2007 11:34:07 -0700, "Saran" <none@nospam> wrote:
>>> cquirke (MVP Windows shell/user) wrote:
>>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:
>>>> A new OS is about scalability, not "speed".
>>> Yes, it's about scalability, but that's not JUST what it's about.
>>> Vista (as Longhorn) originally promised speed as well, and so far
>>> hasn't come close to delivering. Hell, it hasn't really delivered on
>>> the scalability either.
>> Now that you mention it, Windows 4 was going to deliver speed when it
>> was first announced, too. We expected it, as on paper, 32-bit flat
>> memory addressing and no separate DOS layer sounded way more
>> efficient... but by the time Windows 4 finally blinked into the light
>> of day as NT 3.1, it was a resource-heavy slug that needed four times
>> the RAM just to break even in speed.
>
> This is true, but it seems this also shows Microsoft hasn't learned much
> from those days. Also, NT was never marked for home use like Vista is
> for the most part is. It was for business use and business setups
> typically use more resources depending on the type of work than home
> users.
>
>> I guess when Vista was still young enough to be called Longhorn, such
>> claims may have been made, too. I dunno if I don't remember them
>> because I wasn't interested in it at that time, or whether I heard
>> them and rejected such claims before they reached my memory ;-)
>
> Again, it appears Microsoft has learned little in the grand scheme of
> things, other than how to make more money with inferior but flashy
> goods. Its not so different from how some auto makers routinely cut
> corners in production and generally crank out cars that are inferior
> than previous models.
>
>> In practice, Vista does have a lot of speed-up technologyy, but this
>> serves only to soften the body-blows from its huge overheads, much as
>> Windows 3.1's aggressive write caching (with all the file system
>> corruption that implies) merely help it be slightly less grossly
>> slower than DOS as an application platform.
>
> Its the huge overhead in Vista and many new applications that's the
> problem. Too much over head and too little new/gain. That sort of
> equation doesn't work in general cost benefit analysis, so why on earth
> should Vista be any different? Too much cost, in terms of money for the
> OS itself plus the hardware needed just to run it adequately and time
> spend dealing with all the issues as well as the extra overhead itself,
> and too little benefit over previous versions.
>
>>>> Do 1998 programs run faster on a 1998 PC with 64M RAM running
>>>> Windows 98, or Windows 2000?
>>> In my experience most ran better on 2000, unless you're talking about
>>> Win 2000 prereleases.
>> No; thinking Win2000 RTM and SP* - but, in only 64M RAM. I might
>> agree "better" in stability and uptime terms, though.
>
> It depends on what you're trying to run. Normal applications ran so much
> better. The improved memory management over 9x was worth it alone for
> me, and that also translated to far fewer reboots needed (since I left
> 9x years ago I hardly ever need to reboot any system I have.) If you
> mean games, the initial versions of 2000 didn't fair too well but was
> fixed later. I never saw any such problems in XP though, though that's
> not to say there weren't any, just nowhere near the frequency of 2000's
> initial release or Vista for that matter.
>
> It seems to me that going from 2000 to XP, Microsoft really seemed to be
> showing they learned a few things and the response was relatively
> positive. Vista seems like they took a step back in many ways with all
> the many issues. Why break so many things when they were working
>
>>>> Now try the same test on a PC with 1G RAM.
>>>>
>>>> Now try the same test with a 2008 program that requires 512M RAM to
>>>> operate, as well as > 137G HD space. How good is Win98 now?
>>> Thats part of the problem as well; too many programs are far over
>>> weight in regards to the resources they require.
>> Hardware (made in China) is cheap, whereas software (made at least
>> partly in USA) is not - and this is more than a labor wage thingl;
>> there's a genuine complexity factor, too.
>
> So why should one pay high prices when there isn't much new/benefit over
> the previous versions? It's again the upgrading for the sake of
> upgrading and it's just worse when they force you to upgrade (think AV
> and Tax type software.)
>
>> So the trend is to throw hardware at speeding up software, rather than
>> boost software development efforts to better fit smaller hardware.
>
> And you see nothing wrong with that?
>
>> Yup. Software vendors make software, and thus money.
>> They ain't gonna get rich making less software ;-)
>
> Too many make the same software over and over and pitch it as something
> new when it's essentially the same old thing. Some new version are even
> worse, changing what made them so great. ACDSee used to be great, for
> example. I'm tired of being told I need to upgrade or I'm gonna "lose
> out" when a previous version does the job perfectly well. Why real
> service does it to for you to install a new version that's eating more
> resources and doing nothing real new. It just has a face lift, and maybe
> a rearranged GUI, which, once again, is not necessarily a good thing.
> This is also an issue in Vista: make changes for the sake of making
> changes, changing things like labels that have been a particular way for
> well over a decade.
>
>>> What reason is there for this? (I suspect that newer dev kits
>>> being partly to blamed; the .Net frame works, for example, are
>>> seriously bloated. I coudl write two identicle prgorams in VC++ 6 or
>>> lower and in VS 2003 or 2005 and the former would be infinately
>>> smaller. Theres just no reason for so much extra fat.)
>> Actually, there is, if this "fat" formalizes the layers and surfaces
>> between software components so as to reduce bugs and exploits.
>
> But it's introduced even more bugs and there already been news about new
> exploits.
>
>> "Never code anything bigger than your own head"
>>
>> Once a project is too big to fit in a single dev's head, the era of
>> "artistic programming" (as in "I can't understand a word of the source
>> code, but wow, the object code kicks butt!") is over.
>>
>> Attention turns instead to formalizing the way different devs see and
>> (re-)use each other's work, in an effort to kill bugs. The process of
>> killing these bugs is rarely over by the time the product ships
>> (currently, we expect better from hardware development, which is why
>> we are more confident in large hardware carrying the load).
>>
>> If the kill-the-bugs phase is never over, the optimize-for-speed
>> process never even begins. In fact. there's so much fear of
>> regression bugs (new bugs introduced by "improving" the code) that it
>> may be rare to re-cast working code just to improve efficiency.
>
> Well, I partly agree, but I see too many broken things in Vista for that
> to really be the reality as it stands. What I see is (possibly with
> every new release of VS and/or .Net Framework) things just get bigger
> and fatter and slower for little gain between versions.
>
>>>> Vista may run software faster than XP, in the context of mid-Vista
>>>> hardware (e.g. multi-core processors, 2-4G RAM, etc.). Vista's not
>>>> built for 2007 and software of that vintage; it has to support
>>>> several years of future development, which XP is unlikely to do as
>>>> well.
>>> I disagree. The original promises of then Longhorn, now Vista, were
>>> that it weould be faster than XP and previous versions, would be
>>> able ot run almost anything. Instead we have a lardass of an OS
>>> laden with DRM and a mountain of compatibility issues, making it a
>>> far cry from it's promises.
>> I don't think I ever took "it will be faster" claims seriously, if
>> indeed I ever heard them. However, I do expect to see 2009-era
>> software that works better on Vista than XP, and 2010 software running
>> faster on Vista64 than any XP, and 2112 software that requires Vista
>> and won't run on XP at all.
>
> Maybe, maybe not. It's a lot of work to rework applications that
> depended on 32 bit addressing. You can't necessarily just load up the
> source on a 64 bit machine and hit compile. you have to take care with
> data structures that used 32 bit types as recompiling doubles that
> width, and thus takes up more space. Actually, it seems that this would
> cause many programs to eat even more resources.
>
> At some point you have to ask yourself something? Do you want faster,
> more responsive software with a decently small footprint. Do we really
> NEED all these new versions if the previous gets the job done better,
> which a lot of the time seems to be the case these days. IMHO, I'd take
> speed, efficiency and ease of use over eye candy any day.
>
>> The time frames may change, driven largely by external events; malware
>> evolution and to what extend DRM-riddled media becomes pervasive.
>
> I think one could say all that DRM is actually fueling the piracy that
> it want to stop. It seems to me torrent downloading of movies and TV
> shows is at an all time high, and if that's so, I wouldn't be surprised.
> I personally would rather have freedom than be shackled.
>
>>> Then again, many classic dos games (Doom, Warcraft, etc) have
>>> had Windows versions for some time
>> Ain't the same, sometimes. Wolfenstein isn't the same without the
>> flat ceilings, lurid primary colors, chintzy FM music synthesis and
>> gritty 8-bit sound samples ;-)
>
> Agreed. I enjoy the orignal everyone once in a blue moon via DosBox, as
> it runs flawlessly there under XP at least.
>
>>> simpler games could be run under "DosBox" as well, which gives you
>>> far more features the stock cmd.exe ever could.
>> Yep, that's what I had in mind. Eventually, sub-PC devices may get
>> fast enough to port the lot over to a PC emulator on a cell phone.
>
> I would like to see that
>
>>> In the end, I find completely inexcusable that an OS as new as
>>> Vista, on 3+ gz dual core processor to be so slow.
>> It's not that far outside my expectations, especially if you're
>> looking at 5-year OS cycles instead of 3-year NT or 2-year Win9x.
>
> From what I understood, Longhorn goes back at least as far as 2001, so I
> estimate a 6+ year overall development. Combined with the fact they
> already had three rather well working OSes in 2000, XP, and 2003, yet
> Vista does many things all that much worse. With Vista it seems to keep
> coming back to upgrading for the sake of upgrading. 98 to 2000, 2000 to
> XP, offered truly new and useful advancements. What does Vista really
> offer that is really NEW ?
>
>> XP on 2002 PCs was pretty heavy, too, as was Win3.1 on the 2M or 4M
>> 386 systems of their day.
>
> I never had any trouble running Win 3.1 on a 386/33 with 4M of ram
>
> As for XP... it seems to run decently on a variety of systems. I run it
> on a P2 laptop with 400M ram using SP2 and it's incredibly solid. Gaming
> and high res videos aside, in general applications it's almost on par
> with a well oiled p4 I also having running XP. At one time, as an
> experiment, I even installed XP SP1 on P1 200 with 128M ram, and it ran
> surprisingly well. Sluggish? Of course, but stable as hell, while the
> initial startup was nothing to write home about, it still ran relatively
> smooth and consistent, even if loading applications was rather slow. It
> still worked and was surprisingly usable.
>
> Can the same be said about Vista, in the relative sense? Hell, you need
> the new high horse power just to make it run well. I'm still at a lose
> as to why Vista on a decent 3+ GHz P4 or Dual Core almost feels like XP
> on that P2 laptop? Again, this feels more like a step back than true
> progress.
>
>> XP came out as we moved from PIII to P4, Win98 from Pentium to PII,
>> Win95 from 486DXn to Pentium, and Win3.1 from 386 to 486. In each
>> case, you were caught between the exhausted tail of an old hardware
>> platform, and the feeble first of the next generation.
>
> 386 to 486 to early Pentium was never so bad. There was some speed
> difference (and more ram helped too.) Clock speeds seemed to jump around
> 30~ MHz at a time and gave decent gains. When then, with speeds in
> excess of 100 fold (clock wise) from a 33 MHz 386 to a 3+ GHz
> P4/DualCore/etc, performs so slow in comparison. What happened to all
> the gain? 100 times faster in clock speed. 15 times that of a 200 MHz P1
> or 6 times that of a 500 MHz P3. Yes, I know FSB speeds have been
> struggling to catch up, but still, for such clock speed increases I
> expect a lot more out of an OS that's supposed to be so advanced, in
> that respect.
>
>>> I cna compare it to an old pentium 333 mhz 2 laptop I have
>>> laying running Win 2000 that can run most applications about
>>> as well as a typical Vista machine I've used in any shop
>> Yup. I had a Pentium-200 running Win95 "build 950" in 8M RAM, and the
>> old thing rocked surprisingly well. To me, that was the breakthrough
>> OS, just as the original 386 was the breakthrough processor.
>>
>>> Going from a 386 to a 486 was quite a boost...
>> Not really, no - in fact, initially the non-Intel wolves would rip the
>> flesh off Intel's initial overpriced offerings. AMD 386DX-40 or Intel
>> 486SX-25? AMD 586DX4-133 or Pentium-75?
>>
>> But once the 486 discovered core multipliers to become the DX4, and
>> Pentium routinely enjoyed 66MHz base speed and lower operating
>> voltages, that's when things started to rock. Expect n-core
>> processors (and an OS that knows what to do with them), etc.
>
> The problem was that AMD wasn't so well known yet. Intel was the defacto
> at the time, and yes the 486 DX2 66 w/ 8M ram (which is what I had back
> in the day) kicked ass
>
>>> so was a 486 to a P166 or 200
>> That's exactly what I mean. When the 486 hadn't discovered core speed
>> multiplication or sub-5v power yet, the Pentium SUCKED
>
> Agreed.
>
>> (do you remember the appalling P60 and P66?).
>
> *Shudder* I try not to
>
>> When the Pentium Pro tried dedicated 32-bit design for the first time,
>> it was effective, but niched due to costly manufacture and poor 16-bit
>> speed. In fact, it took a while before Intel shrunk the silicon
>> enough to fit the whole processor core in a single die, and it was
>> only then that the PII/III really lived up to its potential.
>>
>> So we've seen processor generation false starts before, even with
>> fairly sub-revolutionary generations (think PIII Slot One vs. S370,
>> and S426 ? vs. S478 P4). We may yet see the same thing with the Core
>> 2 generation, though this time it looks good from Day 1.
>
>
> Yes, but I really don't see how this really changes the facts about the
> current state of the software industry cranking new resource hungry
> version don't give much benefit over the version that came before.
> Again, I tend to place Vista in this category based on the overhead,
> changes for change sake, DRM, etc.
>
> I for one don't like being treated as a sheep by vendors who expect me
> to buy every new release just because it's new. If there's some real
> benefit then it may be worth getting it and it's fairly priced, then it
> makes sense. But as long as vendors (including Microsoft) release
> over-priced, over-weight software with out giving some really benefit
> that makes it worth it, then I'm not going to buy it. Simple as that.
>
> -saran
>
>


I salvaged a couple of old Compaq 350 MHz machines for use at work and
both run XP quite well. I did add some RAM, they were poorly equipped
anyway even for the original W98 that was on there, and with XP RAM is
the key.

On Sat, 11 Aug 2007 14:09:02 -0700, "Saran" <none@nospam> wrote:
>cquirke (MVP Windows shell/user) wrote:
>> On Thu, 9 Aug 2007 11:34:07 -0700, "Saran" <none@nospam> wrote:
>>> cquirke (MVP Windows shell/user) wrote:
>>>> On Sat, 28 Jul 2007 09:33:55 -0700, "Saran" <none@nospam> wrote:

>>>> A new OS is about scalability, not "speed".
>>> Yes, it's about scalability, but that's not JUST what it's about.

>>> Vista (as Longhorn) originally promised speed as well

>> Now that you mention it, Windows 4 was going to deliver speed when it
>> was first announced, too. We expected it, as on paper, 32-bit flat
>> memory addressing and no separate DOS layer sounded way more
>> efficient... but by the time Windows 4 finally blinked into the light
>> of day as NT 3.1, it was a resource-heavy slug
>
>NT was never marked for home use like Vista is for the most part is.

Windows 4 was; it was only when (during the development process) that
32-bit DOS-less Windows was discovered to be such a performance
disaster, that it was re-purposed as a hi-end, stability OS (and the
whole "server" thing took on more significance).

At the time, what happened was; the hype about 32-bit DOS-less Windows
faded away, then Windows 4 development took longer than expected and
non-one much minded because all the "it's coming!" hype had died away.

So instead of a 32-bit Windows 4 to replace Windows 3.x for everyone,
we crawled on through point revisions of Windows 3.x (with Windows for
Workgroups 3.11 being the unexpectedly great "sleeper" version, that
era's equivalent of Win95 SR2 or XP SP2).

During this time, Windows 4 became NT and was done *properly*,
creating a pretty solid, no-compromises, dedicated 32-bit code base
that made no apologies for being a resource-heavy pig that wouldn't
support your current DOS games and dodgy Win16 apps.

And so we saw the two-track Windows, with NT on the high-end "pro"
workstation and server market, and Windows 3.x on DOS for the rest of
us. Hardware got better and MS's ambitions became more realistic; the
result was the "compromise OS", Win9x, which served us very well for
the next... hey, only 7 years; XP's been around for nearly as long!

>It was for business use and business setups typically use more
>resources depending on the type of work than home users.

They didn't so much in those days; they grew to, as business slowly
(too slowly, for MS's taste) switched from Win9x to NT. Part of the
attraction for big business was the user management brokered between
domain servers and NT clients; that in itself required more muscle on
the desktop, what with NT(FS)'s "am I allowed to do that?" fiddling.

>> In practice, Vista does have a lot of speed-up technologyy, but this
>> serves only to soften the body-blows from its huge overheads, much as
>> Windows 3.1's aggressive write caching (with all the file system
>> corruption that implies) merely help it be slightly less grossly
>> slower than DOS as an application platform.

>Its the huge overhead in Vista and many new applications that's the
>problem. Too much over head and too little new/gain.

Now you're getting to the crux of it.

PCI and the NT code base took a while to grow into, but offered
genuine extra value that made for a longer-running future. Will the
extra beef in Vista lay down a similar foundation, or is it just fluff
we don't need? Like you, I may lean towards the latter opinion right
now, but time will tell.

>Too much cost, in terms of money for the OS itself plus the
>hardware needed just to run it adequately

The pain of today's perfornace sluggishness is temporary (tho none the
less unwelcome and inescapable right now).

During the DOS era, PCs were always R 2000 tp R 2500 here; Hercules
mono graphics, 1M RAM, and whatever the processor and HD of the day
might have been. Windows 3.x meant extra software cost and hardware
that blew the cost of a PC out beyond R 4000, where it has grown to
between R 5000 and upwards ever since (it's back in that zone now).

The point is, it costs pretty much the same to make a 1G RAM DIMM as
it does to make a 1M RAM SIMM, once the dev and tool-up costs are
recovered. I've just quoted on upgrading an XP PC where the comparing
the component costs look like this...
- RAM: original 128M = R 1 000, new 512M RAM = R 550
- HD: original 80G = R 1 800, new 320G = R 950
- Optical: original CDRW = R 950, new DVDRW = R 400

See the trend? That's on a 2002 XP PC, and as you can see, it was
more painful to build a 128M RAM PC then than it is to build a 1G RAM
PC today (as 1G DDR2 = R 500 now).

So I see the hardware issues as yap, yap, yap. New sware on old PCs
has always sucked, and that has and will not change... the only folks
who want that, are the software vendors selling the new software.

>>> In my experience most ran better on 2000
>
>The improved memory management over 9x was worth it alone

Well, there's a case in point; it's that memory management that scales
up to the "large" RAM we have used throughout the XP era. It may not
have seemed relevant when 64M was all the market could afford...

>that also translated to far fewer reboots needed

That was the resource heap thing, which is making a comeback in XP and
Vista right now. This is where you could insert "MS didn't learn
anything", given that we're still faced with a finite global store vs.
per-instance data of an unbounded number of instances.

But at least the fix is a registry setting, whereas with the old
16-bit heaps, it was more like waiting for the earth to change.

>If you mean games, the initial versions of 2000 didn't fair too well
>but was fixed later.

Ahh, "fixed later"...

>I never saw any such problems in XP though

That's because XP was built on Win2000's foundation, even though that
foundation was painfully heavy (too heavy for the consumer market)
when it emerged in and failed to take over the Win9x market.

Vista is setting a new foundation - but will it be as solid and
long-lasting? Will it, too, be "fixed later", and by 2009 be the only
sensible and scalable way forward for Windows?

>It seems to me that going from 2000 to XP, Microsoft really seemed to be
>showing they learned a few things and the response was relatively
>positive. Vista seems like they took a step back in many ways with all
>the many issues. Why break so many things when they were working

I think when you look at the spam statistics, and the number of PCs
that are running malware (often where the user would rather tolerate
that than submit to "just wipe and rebuild" because we no longer have
any other end-user solutions), one may question how well things are
(and will continue to be) "working".

XP SP2 is much better at meeting today's challenges, but in time
scales of a decade or more, it may be looked back on as we do Windows
for Workgroups 3.11 - sure, it's a better runty crash-o-matic OS (or
in XP's case, food trough for botnets) but you can't seriously expect
us to build the next decade's computing on *that*.

>So why should one pay high prices when there isn't much new/benefit over
>the previous versions? It's again the upgrading for the sake of
>upgrading and it's just worse when they force you to upgrade

Right now, you shouldn't; if you have a working XP PC, stay with that
awhile instead. But build a new XP PC right now; nah, I wouldn't do
that, either. In fact, seeing as 1G RAM now costs about what 512M
cost at the start of the year, a Vista PC isn't really that much of an
overspend anymore.

>> So the trend is to throw hardware at speeding up software, rather than
>> boost software development efforts to better fit smaller hardware.

>And you see nothing wrong with that?

I don't, no; even as the old programming artist within me writhes in
discomfort, I recognise that today's challenge is no longer scaling up
the PC to fit what we are trying to do, but strengthen it's guts so it
doesn't collapse under the weight of its own complexity.

>Too many make the same software over and over and pitch it as something
>new when it's essentially the same old thing.

Yup. The industry is mature, or even post-mature.

Telling signs are the changing ratio between efforts to make new
things possible, vs. the effort to make currently possible things
artificially impossible. It's like watching the Industrial Age
striving to protect the jobs of hand weavers, horse farriers and slave
ship companies against the natural applications of technology.

The US economy has become a mix of brand hysteria and hyped
entertainment. They literally don't "make" things anymore, though
they do a lot of applied psychology, design, and genuine invention.

>>> What reason is there for this? (I suspect that newer dev kits
>>> being partly to blamed; the .Net frame works, for example, are
>>> seriously bloated. I coudl write two identicle prgorams in VC++ 6 or
>>> lower and in VS 2003 or 2005 and the former would be infinately
>>> smaller. Theres just no reason for so much extra fat.)

Which is the biggest challenge for a modern program, over (say) a
3-year time frame:
- that it will run too slowly on the hardware of the day?
- that it will crash, or be exploited by malware?

The answer to that, is why the bloat of simplified, shallow-depth,
layered tools is embraced by today's progerammers.

C++ is a "deep" tool, allowing you to structure the top design level
of the code via classes etc. and yet drill down to the details of what
values get stored in what registrers. How many folks can hold all of
that in their head? How many folks have the skills to both lay out a
well-designed set of classes, and fiddle around with byte type
overrides without screwing up somewhere?

It's funny, though, because I remember when C++ was seen as the
Devil's Spawn of software bloat. "This program's so big, it won't
even fit on a diskette! I could have coded it in half the size in
assembler or straight C without those bloaty objects and libraries!"

>> Actually, there is, if this "fat" formalizes the layers and surfaces
>> between software components so as to reduce bugs and exploits.

>But it's introduced even more bugs and there already been news about new
>exploits.

I don't think it has, no. Just as Vista's go-faster technologies have
failed to offset the extra bloat, so modern dev tools' software
engineering safety nets have failed to catch all the leakage.

IOW, if you tried to code, say, Word 2007 in pure C, it would either
be ready for release in 2012, or crash before it finished painting its
window. Whether one would want to code Word 2007 when Word 6 meets
your needs just fine, is another kettle of worms ;-)

In programming, "speed" has come to mean development time, not
execution efficiency. Compare that to the way cars are manufactured
today, or how goods are hauled in containers.. same trade-offs.

On 64-bit computing, you said:

>Actually, it seems that this would cause many programs to eat
>even more resources.

That was an issue with 32-bit computing, as well as Unicode. Folks
really drang the Kool-Aid on 32-bit, decrying Win9x's use of tight and
solid UI code that happened to be 16-bit, which was a core reason why
in ran in 4M RAM the way NT would run in 16M.

But the benefits of 32-bit weren't just goodbye to segment registers.
There was a wealth of memory management (paging, protected spaces,
remapping, the whole hypervisor thing) that drove 32-bit computing.

I'm not sure if that's a factor with 64-bit computing, unless MS digs
in its heels and refuses to carry over compromises from the 32-bit
computing age. The swich to 64-bit breaks compatibility, and MS needs
to grasp that as an opportunity to ditch old mistakes and flaws.

That means switching to 64-bit will take longer to gain momentum, due
to constrained device drivers (how often did you have to click though
"this driver is not signed" in XP?) and software uptake.

>I personally would rather have freedom than be shackled.

MS freed computing for personal use (something we forget today), but
sold its soul when it promoted the "remote administrator" over the
local user in NT's quest for the enterprise. That sale continues,
just to new demons, via DRM and other "trusted Computing" initiatives
(which basically translated to "your malware can trust the system to
disallow the end user to butt in on whatever you want to do").

That's part of what I mean by a "post-mature" industry.

BTW, the old white-coat mainframe brigade never went away; they've
been trying to lure us back to time-share serfdom through the NetPC,
and now Software As A Slavery (sorry, "service"). Put your data on
our big boxes and trust us to secure it! Log into our application
server and do your work here - it's free (for now) !

>What does Vista really offer that is really NEW ?

I'm not sure, to be quite honest. There's what you see, but the real
answer to that will be deeper, and I can't see that deeply anymore.

>> XP on 2002 PCs was pretty heavy, too, as was Win3.1 on the 2M or 4M
>> 386 systems of their day.

>I never had any trouble running Win 3.1 on a 386/33 with 4M of ram

It was slow, and stayed a lot slower than DOS until we saw
hardware-accelerated SVGA. I used Win3.0 and 3.1 in those years on
various 386 systems, evolving towards 386DX-40 and 4M RAM, but that
rig cost far more than I'd ever spent on a PC, and still did most work
in "faster" PICK and DOS with dips into Windows.

My main lure to Windows was word processing, which had previously
meant text editing in a word processor then layout and fonts in
Ventura. Dilbert's "the joy of beautiful documents" was in full
effect - I'd Ventura my grocery lists! - and Windows word processors
like JustWrite that seamlessly merged wp and DTP were cool++

A couple of years on, a 486DX2-66 with VL-bus accelerated SVGA finally
let Windows run as sweetly and fast as its designers may have hoped.

>I even installed XP SP1 on P1 200 with 128M ram, and it ran
>surprisingly well. Sluggish? Of course, but stable as hell

And that's the payoff.

>Can the same be said about Vista, in the relative sense?

I hope so, but right now, it's far from apparent. I hope Vista's
long-term advantages go beyond the opportunity to keep paying movie
rental, as if fat bandwidth and storage had never happened.

>The problem was that AMD wasn't so well known yet. Intel was the defacto
>at the time, and yes the 486 DX2 66 w/ 8M ram (which is what I had back
>in the day) kicked ass

I bought and sold zero Intel throughout that age. All the
best-of-breed processors were AMD; 286-20, 386DX-40, 386SX-40,
486SX-40, 486DX2-80 (<cough> I hated that one), 586DX4-133. All
cheaper than Intel, all out-performing Intel's first offerings in
their over-hyped new platforms.

I have an old IBM 4.77MHz XT motherboard with co-processor, and it is
covered with AMD logic chips. AMD were there from the very start.

Still, they were coasting on Intel's efforts. Once they were forced
to design thier own cores in the Pentium age, the wheels fell off; the
K5-75 was probably THE most useless processor design I have even seen
or wish to see, including even the "intestines in a wheelbarrow" PII
and "floor sweepings for Compaq" no-L2-cache Covington.

AMD recovered by buying out NexGen's team, who had failed the
challenge of establishing an alternative platform, and bringing out
the K6 onwards - ironically, being forced onto their own platform by
Intel's patent and legal maneuverings.

What Intel did, and AMD failed to do, was take control of the
motherboard chipset and set solid standards there. That was the only
reason I stayed with Intel rather than AMD from Sloth One onwards.

>I for one don't like being treated as a sheep by vendors who expect me
>to buy every new release just because it's new.

They won't, if you buy into "software as a service" or "rental
slavery". Sware will cease to be a thing you "own" in perpetuity, and
will be something you never stop paying for, whether you like new
versions or not - and if vendors decide to push a "free" new version
that needs bigger hardware, you'll buy in or fall off the platform.

There are genuine reasons why sware vendors need this; namely, the
swing between pre- and post-release (patch) development. But it's a
change that rewrites the "contract" between user and vendor, and if
today's EUL"A" trickery is anything to go by, we will be soo screwed.



>--------------- ----- ---- --- -- - - -
Dreams are stack dumps of the soul
>--------------- ----- ---- --- -- - - -