On 2 Oct 2000, at 18:57, Robert Deering wrote:
> 'Twas this disk that had me waiting anxiously for the larger release.
> The floppy is a _demo_ of what QNX can do, not a QNX distribution by any
> means.
My mistake. I have been impressed by magazine and website
reviews, but not been bothered to download anything. I have a
friend with a CD burner, and that mirror site is easily downloaded
from, so I might partake.
> Any time a 650MHz processor requires 5 hours to process 4GB of data, no
> matter _what_ the task, something is wrong.
Umm, Decode MPEG2 and re-encode to MPEG4 (aka "DivX ;-)")?
Remember this is a very compressed format being compressed
into an even more compressed format (ie 4GB down to <600MB)
without loosing much quality. One needs at least 500MHz to replay
this video properly (however that is through WMP and theres lots of
overhead there).
> Chances are that most of
> this trip is unnecessary. If the task were set up from boot by a
> programmer dedicated to sensible code, the Duron could do it in 30 or 40
> minutes. I've seen such accomplished before. Properly instructed, my
> 286/287 could probably do it (given enough storage) in about 3 1/2 hrs.
Even if a 650MHz processor took <1 hour to do it, a 20MHz
processor is not exactly (650/20=) 32.5 times slower. There are
such things as multiple pipelines, out-of-order execution, more
instructions (MMX, 3dNow, etc), faster memory/disk access, etc. I
would estimate the Duron would be at least 100 times faster than
your 286. So that means your 3 1/2 hours would be closer to 50.
And would you like to sit and reprogram something that already
exists and optimise for speed? Especially "from boot" as then
you'd need to reimplement all the disk-reading/writing and screen
routines, among others. 5 hours might be a long time, but I don't
see how anyone can shave much time off this.
I'm still not bagging the SurvPC, just saying they're not suitable for
this high-end work.
> My point being that better programming negates the need for super
> machines for most users. Sloppy cut-paste non-programming is _seemingly_
> quicker to do, so it _looks_ more profitable.
I can see both sides of the argument. These programming
practices (which include many Rapid Application Development
packages) are fast to produce, but produce code that has many
parts that shouldn't be included.
Programmers tend to be lazy, and somewhere I was reading about
the open source movement is that they like to share code. Its
"constructive laziness". But many new programmers take all this
and slap it together without really knowing whats going on.
However they don't notice it because they have the latest whizbang
computer, and assume that everyone else has one too. If you don't
there's usually alternatives, but for more of a price. (Cheap
hardware/expensive software or expensive hardware/cheap
software) But I'm not saying all new software is cheap (notibly M$)
but one can get a lot for free (eg StarOffice) that needs 'better'
hardware.
But as many of you have realised, its not econnomically sound for
any company to keep trying to support oldies such as those on
this list. The sheer costs in marketing wouldbe way too much, and
advertising on the web wouldn't produce much results (majority of
websurfers using newer computers). As well as paying
programmers to do it, and Good Old Fashioned Coders are
becoming extinct.
I've been a member since just after this list started, and still use
my 386 and 486, but do most of my real work (and games, etc) on
my new motherboard/CPU.
> Someone's recent post (I
> think it was this list) about so-called "market forces" was right on.
> The majority will never insist on getting real value from their
> machines, but there's no reason for a very large minority (that has
> members on this list) can't do enough insisting to get something done.
> There're many ways to instist. The main one is simply not to take what
> is offered.
That is why everyone is sticking with older versions of software,
designed for the machines of the time. You can't really do that
these days if you are on the internet, because of security
problems. But earlier software (ie most DOSes) didn't have these
problems (not designed for use on the internet!)
> > Even HTML rendering can take a bit of CPU power. Look how slow
> > Arachne is!
>
> There's a lot I'd wish to say here, but I'm using Arachne by the good
> grace of Michael Polak.
I use Arachne too. I wasn't saying it was bad, just demonstrating
plain DOS isn't really as good as some other OSs for this kind of
work. I'll have to check out Arachne for Linux one of these days...
That was quite a ramble. It probably doesn't show my true feeling,
but what the hey.
To unsubscribe from SURVPC send a message to [EMAIL PROTECTED] with
unsubscribe SURVPC in the body of the message.
Also, trim this footer from any quoted replies.
More info can be found at;
http://www.softcon.com/archives/SURVPC.html
