I don't know if it's typical or not - certainly that example is dramatic.
But I do know there are Canon L lenses that are a hell of a lot sharper than
that, so maybe that's the worst of the lot.  For example, the 24-70mm L is
widely considered Canon's best lens period, and many reviews state that it's
as good as any prime throughout its range.  (I happen to be picking up that
lens before the month runs out, while Canon's rebate is still on, so I'll
post my own results soon.)  I happen to own the Canon 70-200mm f4 L, and
though I've had it nearly a year, its sharpness continues to astonish me - I
mean, just astonish me at times.  Sometimes I load up new pictures I've
taken with it, and just sit there with my jaw drooping at how gorgeous some
of them come out, exceeding my expectations even as I was taking the shot.
Mine is the f4, not the f2.8, but I understand the f2.8 is every bit as good
(just faster).  I couldn't recommend it more - it's just that freaking
awesome.  :)

You talk about "probably a Sigma and not a Canon L", but maybe that's the
real difference right there.  The Canon L lenses are tough to beat, and
certainly no Sigma lens is going to approach them, prime or not.

Sure, your style is definitely something to consider.  I'm biased, because I
use a myriad of focal lengths, so primes are simply not for me.

Yes, that's fairly typical of the difference between primes & even very
good zooms. For a real shock, try comparing the difference between a
cheap prime & a consumer zoom. ;)

The same's true of me. I do own a few zooms, but I hardly ever use them,
compared to my primes. If I knew how little I'd use them at the time I
bought them, I would've used the money on a wider range of primes
instead.

Of course a prime lens is got to be sharper than a zoom lens. However,
what you were reading is about how 100-400L sucks. I read another review
that compares this lens to the equivalent Nikon's offering. The Nikon is
sharper and Canon has to stop down to f/11 to be comparable in
sharpness. This is why I got the 300/4L instead of the 100-400L. I have
17-40/4L and 70-200/4L and they are great, so don't worry much about
zoom vs. prime. it's only the 100-400L is less than ideal. I also have
50/1.4 for low light.

Damn, that's a big difference. It's not even fast. Hmm.

It's what I'd expect from one 100-400 zoom vs another 400mm prime lens, it's
not what I would expect from every zoom vs every prime lens. He might have
gotten the same results from a different sample of the same lens and I've
seen shots posted on that site between two lenses that were focused at
different points in the same scene and then compared to each other?

But that zoom doesn't have a very impressive MTF graph so it very well
could be the case. Some of the newest zuiko digital zooms have a better MTF
graph than that prime so I guess it just depends on which zoom lens is
being compared to which prime lens? I'm assuming that 100-400 is from
pre-digital days and was designed with the limits of 35mm film in mind.

Prime or fixed focus lens? What is the difference between a prime lens
and a fixed focus lens?

Alan

It's not only about sharpness; don't forget vignetting and - especially
- distortion. Distortion is what especially turns me off zooms.

There are also minor but meaningful charateristics such as bokeh,
3d-rendering, color saturation and others which distinguish great lenses
from the good. But often these are not noticed or considered important
by many

The basic combination I use is 35mm + 85mm or 105mm, which is a classic.
In my case, other lenses in order of use are: 24mm, 20mm, 180/200mm,
15/17mm.
My 300, 400, fisheye & macro lenses are rarely used. YMMV!

It is. What you should be checking out when shopping for a zoom is its
range of coverage. The smaller, the better the zoom is likely to be
("better" with regard to image quality AND/OR speed).

Yes, that means that those that buy 35-350mm and 50-500mm lenses (10x
zooms) are getting sub-par resolution.

" am looking to move up from a Canon S45 to a SLR.
I was thinking of getting a 70mm-200mm f2.8 Zoom (probably a Sigma not
a
Canon L) until I ran across this:
http://www.luminous-landscape.com/reviews/lenses/forgotten-400.shtml
If this is a typical difference between a zoom and a prime then I think
I
might go for the prime."

You should see what the Leica lenses do before you plunk down money for
lenses!

They trash the competition.

Yes, it is typical. But there are zooms that are almost as good as
primes.

Examples:

* EF 24-70mm 2.8 L
* EF 70-200 2.8 L
* EF 70-200 4.0 L
* EF 17-40 4.0 L

These are roughly equivalent to primes, and if they are not quite as
sharp, the difference is minimal.

Other Canon zooms which do not carry the L-label are not so good.

The 100-400 L is not sharp, and neither is the 35-350 L which I have
tried. The EF 28mm 2.8, EF 35mm 2.0, 50mm 1.4 USM and 100mm 2.8 USM
Macro are all incomparably superior to consumer zooms like the 18-55
3.5-5.6 (does not fit on all cameras) and the 28-135 3.5-5.6 IS. I
know since I own all these primes and have compared them to the
consumer zooms.

The disparity in quality is truly dramatic, but most people are
unaware of the difference until they see or do a direct comparison.
They become satisfied because it looks decent, but the affordable
primes and the L-zooms are just worlds ahead.

Per Inge Oestmoen

I dunno about the performance of the zoom here.  It looks even worse
than the MTF graph suggests.  

I notice that almost every time somebody does a lens test they do not
say how they focused the lens.  The worst possible thing to do - apart
from not focusing at all - is to rely on AF when testing a lens.  That
comparison in the link looks very much to me like the zoom was actually
out of focus.  The blur looks too smooth to be the result of
aberrations.  I would like the tester to state whether he accurately
focused the lens manually, or whether he used AF; and if AF, whether the
focus was actually on the enlarged subject area.  If MF, how are his
eyes - are they able to accommodate the apparent VF distance, or is his
view of the viewfinder not quite sharp?

In fact, using AF when testing a lens will invalidate the test.  AF
relies on DOF to an extent.  It has to accept a degree of unsharpness to
prevent constant hunting, so some degree of hysteresis is built in to
the focusing algorithm, which simply means that if the focus is within
the hysteresis parameter, the camera accepts it as focused.  With 3D
subjects - ordinary scenes, etc. it doesn't matter, but with 2D, or
flat, targets, it can be off enough to produce the results we have seen.

These observations might seem nitpicking, but when one sets out to
evaluate the performance of a lens, rigorous attention to these matters
is essential if the results are to be meaningful.

Last, he tested one example.  'nuff said.

Colin.

PS: I don't have the lens, so I'm not biased here.

Another point no one mentioned, do you plan on always using a large tripod?
If not, you'll NEVER see the difference posted here.