Re: Graphic converter query - transparency

Rowland McDonnell <real-address-in-sig@[EMAIL PROTECTED]
> wrote:

> Jim <jim@[EMAIL PROTECTED]
> wrote:
> 
> > Rowland McDonnell <real-address-in-sig@[EMAIL PROTECTED]
> wrote:
> > 
> > > <very puzzled again>  No, 600+kB/s is around about 338kB/s as far as
> > > handwaving regarding this sort of thing goes.  Double or half the
data
> > > rate is `about the same'; ten times or a tenth isn't.
> > 
> > That's subjective of course.
> 
> <puzzled>  No, it's conventional scientific *****sment of numbers.

I don't care. If I'm only getting half my download bandwidth, that's
significant. I don't care what the overall magnitude is.

>  I'm
> just using the sort of conventional hand-waving judgements that any
> scientifically knowledgeable person would use when faced with the
> variations that exist in reality - in this particular case, I'm looking
> at the variations that exist in typical internet downloads using a
> broadband connection, assuming that what I see is typical (and it is
> typical, from what I've gathered reading here and talking to a few other
> people).

Very nice, but your conclusions are incorrect for me. 


> > To me, double-or-half isn't "about the
> > same" at all, 
> 
> What I say on this subject is indubitably true - although for a given
> context.  What context?  One in which the numbers vary by orders of
> magnitude.  When you're dealing with variations that large, double and
> half are `about the same'.

Except we're not dealing with orders of magnitude, are we? The average
broadband rate in the UK is roughly 3MBps. I personally get about that.
I do not, however, occasionally switch my router off and connect via a
1200/75 modem, which *would* be an order of magnitude different. So if I
start to download rates that mimic that, it's a problem.

> (Do rememeber that I'm the one who wanted to be an experimental
> physicist and did the degree to do so and I was pretty damned good in
> the lab though I say so myself)

Irrelevant.
 
> And that's what we get with internet data rates, innit?
> 
> Since no download I've seen has ever had a rate that's stayed put to
> within 20% while it's been happening, I don't see how you can think
> you've got it right on this one.

I regularly have downloads that max out my download bandwith for the
entire duration of the download, sometimes lasting in excess of several
hours. So you're wrong.
 
> I see download speeds varying from less than 1 kB/s to several hundred
> kB/s, depending.

That's nice. So do I from time to time (depending on what I'm
downloading, where from, and what the time of day is) but I also see
downloads that make full use of my bandwidth.
 
> Your +/-20% band for `about the same' is clearly not quite right.

That's your opinion and you're welcome to it. As I said to begin with,
it's subjective. To me +/- 20% is about the same for a typical (for me)
download. For someone else it might not be.

>  I'd
> say that if you had two downloads, and they arrived at an average data
> rate within 20% of each other, that'd count as `practically identical
> data rates' from the `broadband internet download' point of view.

Yes, for me. Someone else might want to investigate where the bottleneck
was.

> 
> >it's significantly different. +/-20% would be "about the
> > same" to me (where typical broadband speeds are concerned), but no
more
> > than that.
> 
> That might be your particular point of view, but it's not the general
> one and it's certainly inappropriate for this kind of thing.

I don't think you know what "the general one" is.

Jim
-- 
'Cloverfield' in nine words: "What is it?!" "We're gonna die!" BOOM!
Roll credits.

http://www.ursaminorbeta.co.uk
  http://twitter.com/greyareauk