Posted by dorayme on 12/20/07 21:15

In article <5sv93dF1afqg9U1@mid.individual.net>,
Bergamot <bergamot@visi.com> wrote:

> dorayme wrote:
> >
> > It is not wise to be loading anything like 1.5MB onto a web page.
> > One would need a screen the size of a wall to see it.
>
> Um,

Let me say it again then:

It is not wise - and there is no uming about it - to be loading
anything like 1.5MB onto a web page.

Let me quote myself a little from the post you reply to:

"There are a number of things you can do to get the
thing to be suitable for the web. One of them is not to go bigger
than anyone can see on their screens. ... Biggest *surprise* to
folks should be about 800 wide, 600 high. That is one thing, a
big determinate of file size but only one.

There is then the question of how to prepare or compress...
It is not wise to be loading anything like 1.5MB onto a web page.
What possible benefits could there be. One would need a screen
the size of a wall to see it."

(The *last sentence* was an exaggeration, of course. I was
thinking of some pictures I and others have prepared that look so
fine at 1000px across that are no more than 180px (but often much
lower) and that to maintain similar quality and show at 1.5MB at
100% would require quite a big screen)

>there is no relationship between the dimensions of an image and
>the number of bytes.

This, of course, is *almost perfectly* untrue in practice. It is
even perfectly untrue full stop.

If indeed there is no relationship between the dimensions of an
image and the number of bytes, then it should perhaps be
considered quite unusual or coincidental that people the world
over experience vastly reduced byte size of their image files
with little normal loss of quality by simply reducing the width
and height and resaving (especially if the master file worked on
is lossless)

> I can make a full screen size image that's less than
> 2KB. Likewise, a 100x200px image can be several hundred KB in
> size.

Smallest I can do for my biggest screen at 1600 by 1200 is 0.4KB.
But then it is not much of a "picture". Still an image file!

But hey, let us not fight on examples and practical day to day
grounds. That would be altogether too boring. Let us look at this
from the purest point of view.

Let us look at a graph of all the ordered pairs of images against
file size. This means specifically constructing as a prelude to
graphing, a simple 5 col table, 2 cols grouped to represent the
before and after size of images. 2 more cols to say the before
and after of byte size. And last col to give the percentage
difference in byte reduction.

What images shall we put in? Why, of course, every single pair
that has ever been worked on by anyone in internet history. That
would be a fair test, no? And it would give the lie to your
statement that there was no relationship between the dimensions
of an image and the number of bytes.

--
dorayme

[Back to original message]