Strongly suggest you go to dpreview.com and then on to Steve's Digicams
for reviews. These are not totally objective, but, then, nothing is,
and they're far better than anything else around. There is enough
information supplied that you can draw your own conclusions fairly
easily on the camera bodies.

Lenses are a problem right now. No one seems to be doing reliable use
testing of particular lenses. The old adage, "Ask the man who owns one"
might be helpful here. DPreview has forums on different camera brands
that offer a reasonably high level of information without as much ax
grinding as you may find here.

Here's a gallery of Lacrosse pictures taken with the 350D and the 70-200f4L
lens. If you click on original, you can see the full size image as it came
from the camera. Judge for yourself.

I also use the Canon 100mm macro USM lens. Check out the "Up Close and
Personal" gallery and "Flowers and Plants" for that one.

-- Martin

The short version:
I recommend you get the camera you like, decent lenses, and learn the zen of
raw image processing in PhotoshopCS2 (it has a nifty filter for correcting
distortion too!).

The abbreviated long version:
The more hours I spend looking at freshly opened raw images in Photoshop the
more I believe that the comparison between lenses used in a film system and
in a digital system is not straightforward and may be irrelevant.
If you enter digital photography the quality of the lens, while still
extremely important, is not as absolute a value as with film based
photography. The sensor itself and processing are as important or more so
than the lens once you reach a certain minimum quality of lens, which is not
all that great. You can change film but you cannot change the sensor . . .
Inescapable software image processing, whether in-camera or later
in-computer, significantly affects what comes out of the digital imaging
sensor as much or more than what the lens produces on the sensor.
As of now the resolution/noise of digital sensors generally lags behind what
good, let alone great glass can produce. As importantly the physical
interaction of a digital sensor with light is different than the physical
interaction of film with light.
"Sharpness" and "acuity" are far more relative terms in the digital world
than in the film world because most people confuse actual sharpness or
acuity with what they perceive they are seeing on their computer monitor
without taking into account the extreme amount of computer processing that
goes into the rendering of every digital image.
Every digital original image is "soft" and even a freshly opened raw image
has had some sharpening algorithm applied to it before it hits the computer
monitor.
In the real world most of us have very little rational basis to define the
absolute sharpness of a lens using a digital camera, at best you can only
compare images made by different lenses with as little processing as
possible applied to the images. In the real world, as opposed to the
imaginary world of blown up parts of images on a computer monitor, you
really do not know how sharp your image is until you print it--and printing
is impossible without software sharpening!
Issues like chromatic aberration are more significant  to a digital sensor,
partly because of the way the sensor is sensitive to the angle at which it
is struck by light, than with film but are also to a degree correctable with
software. Chromatic aberration from a wide angle lens that might not be
visible in a film image can create shimmering edge effects on a digital
sensor in high contrast areas. This is a very visibile problem if you want
to use wide angle lenses with a digital SLR (you are down in the 12-18mm
focal length range where even things like light diffraction by the diaphragm
can affect the image).
If you learn to recognize barrel/pincushion distortion in your lens and have
a software filter that effectively neutralizes it does your lens actually
have that distortion? Does your lens have chromatic aberration if it is
removed by software such that it is no longer seen on your monitor or in a
print?

That's a given.

You probably would be happier with the 20D body.

The E-1, and other Olympus dSLR cameras, be avoided.

You really can narrow down your choices to Nikon or Canon pretty
quickly, since they're the only ones with even close to a full line of
high end lenses. With Nikon, wait for the D100 replacement that should
be out soon.

See http://digitalslrinfo.com for more information. It's the premier web
site for digital SLR information. Or google "digitalslr unbiased info"
and it'll be the second hit in the list (the first that isn't a domain
for sale!). ....and I'd say this even if it wasn't my site!

You will.

I can only speak of the olympus lenses you mention but they are both as
sharp or sharper than any primes I've ever used.

Either should work fine depending on what you shoot. If high ISO shooting in
low light is something that's important, go with the canon. I personally
like the look of the images and colors from the olympus better but that's
just my subjective opinion. That 50-200 is a fantastic piece of glass with
wonderful bokeh.

http://img.villagephotos.com/p/2005-1/937049/couple.jpg

http://img.villagephotos.com/p/2005-1/937049/roseweb2.jpg

Each system has strong and weak point so you need to look at the good and
bad points of both systems and decide what's important to you. I liked the
idea of self cleaning sensor, slightly deeper DOF 4/3 has at the same
fstop, spot metering and several other features so went with the E300 over
the canon. I'm sure other may have different needs like a deeper lens
selection, more AF points, avalability of more used lenses, the need for
fast prime lenses and would decide the canon fits their needs better.

I doubt it. I bought the E300 with it's 14-45 lens and tried the 40-150 and
while they are pretty good, the 14-54 and especially the 50-200 are MUCH
better. Once you've used good glass (which is looks like you are using now)
you won't be happy with the cheap consumer zooms.

Best thing is to go somewhere you can try them out. Hopefully they will let
you shoot a few shots with each, have them printed at their lab and look at
the results first hand and decide which looks better to you. IMHO what the
images look like blown to 100% on screen doesn't metter to me, what the
final print looks like is what's important.