Text 20747, 79 rader
Skriven 2006-09-08 19:02:00 av CHARLES ANGELICH (1:123/140)
     Kommentar till en text av ROBERT FOWLER
Ärende: Re: slowdown of late
============================
123ecb4e7a2d
win95



Hello Robert - 

CA>>> I mis-spoke or over empahasized the use of "interesting".
CA>>> Honestly I'm not so sure what people here find
CA>>> interesting that pertains to Windows. 

JB>> Ah... Making it run half as well as the competition? 

CA>> The "run half as well" is a bit vague. Give me three
CA>> specific tasks that an OS you own/use can do at twice what
CA>> a recent version of Windows (98SE forward) can do using
CA>> the same or comparable hardware. 

RF> Actually, it's a well known fact that the tools design
RF> architecture of .nx based OSes, which includes both MAC
RF> OS-X and Linux, is much simpler and less bloated than the
RF> application design architecture of Windows. 

That may have been true at one time but I'm doubting that is
true now in 2006. 

RF> Other things being equal, leaner code performs better on
RF> comparable hardware. 

Not really. The better algorithms outperform those with a
lesser efficiency. A significant portion of code remains unused
until it is required for whatever the program must produce.
Code to create multiple versions of MP3, OGG, AAC, and others
exist within an application but they are not all equally
utilized (normally). Those extra functions bloat the size of
the executable but in no way reduce the efficiency. 

CA>> I find that ISP bandwidth, hard drive transfer rates, and
CA>> FSB rates are a problem and those would affect _every_ OS,
CA>> not just Windows. 

RF> The point is well taken, although I would note that ISP
RF> bandwidth is a relative non-factor. Network bandwidth,
RF> including wireless, wired, USB and Firewire, etc., is
RF> another matter. 

I don't know the _actual_ bandwidth for the internet backbone
but it isn't slow by any normal definition of the term slow.
ISPs throttle users down repeatedly until they begin losing
customers then they open up the 'pipe' for awhile and start
throttling again. :-( 

RF> Still, the leaner code should have the advantage, IMO. 

Somewhat but only in time-critical applications that attempt to
cope with real time. In a driver I produced to emulate a UART I
counted nano-seconds per instruction and replaced the ones that
required too many of them. This was assembly language coding,
not general application coding with a user interface. 

Again, size does not always indicate efficiency nor a lack of
it. 

RF> The obvious discrepancy between Windows and other
RF> platforms, is that there is much written solely for
RF> Windows. Comparisons at that juncture are meaningless. 

For the less-than-technical that is a MAJOR factor. :-) 

>
>        ,                          ,
>      o/      Charles.Angelich      \o       ,
>     <|                              |>  __o/
>     / >          USA, MI           < \   __\__
 

--- * ATP/16bit 2.31 * 
... DOS the Ghost in the Machine! http://www.devedia.com/dosghost/
 * Origin: Try Our Web Based QWK: DOCSPLACE.ORG (1:123/140)