Thomas:
   As Ed has told me from time to time  --  "Get your attributes straight"
   Your first line above "The 12-24 DX lens is supposed------------ body"
is certainly NOT mine.  I have no idea how the 12-24 DX lens would work on a
film body and I rarely comment on anything I haven't used or don't own.  I
have read, however, similar comments which you have quoted.
Bob

Heck, how much do you want to pay now? I'll mark one up as high as you want
today! No need to wait.
--
Ed Ruf    Lifetime AMA# 344007 (Usenet2@EdwardG.Ruf.com)
http://EdwardGRuf.com

The cropped sensors work well for people into things like wildlife
photography were for a given f.l. lens you can put more megapixels
in a given area than a larger sensor camera since most wildlife shots
end up cropped to some degree.  A smaller 300mm lens becomes
in essense a 450mm lens.  Less weight, faster reaction time, etc.
The camera is also likely to be faster when it comes to things like
rapid shooting, 8 fps, and so on.
Canon's MkIIn 8 meg. exists as a $4000 body because of this.  Action
photography, etc.  However, for landscapes other areas where wide
shots tend to predominate, the 5D seems more popular, despite its
interaction with Canon WA lenses. It also (from what I've seen)
appears to produce more natural looking portrait shots than the
cropped cameras.
-Rich

The P&S cameras typically lie and give "35mm equivalent" focal lengths to
confuse you.  So that would be "35mm equivalent".  The "35mm equivalent"
field of view of 18mm on DX would be 27mm.

As Jeremy noted this is the 35mm equiv. Look at the front of the camera. It
should be inscribed with the actual fl: 6.1 - 26.1 mm
--
Ed Ruf    Lifetime AMA# 344007 (Usenet2@EdwardG.Ruf.com)
http://EdwardGRuf.com

No, I didn't!  This is a complication; I was just starting to get
comfortable having made a decision, and now I discover my reasoning
was flawed.  How inconvenient :-(

In a conventional lens, the central rays of each bundle forming a point
in the image diverge as you go from the centre of the frame towards the
edge and the corners.  You can get a rough idea of how much this is just
by working out the actual diameter of the lens necessary for the focal
length at the maximum f/#.  For example a 20mm f/2 lens will only be
10mm diameter at its stop, however, the diameter of a 35mm frame is
around 43mm, so somehow those rays have to reach from the edge of the
lens out to the corners of the frame, and that means the central ray in
each bundle must be leaving the lens and arriving at the focal plane at
a steeper angle than those closer to the centre of the frame where the
central ray is almost perpendicular to the frame.

The main effect of this is that when images are out of focus they can be
smaller or larger than the image when it comes into focus, and you may
well have seen small effects like this already.  Obviously for
measurement systems, such as machine vision, you don't want small focus
errors affecting the sizes you are measuring, so simple conventional
lenses are inappropriate for that type of application.

On a telecentric lens, the lens design ensures that all of the central
rays on each bundle arrive at the focal plane perpendicular to it and
all of the peripheral rays are evenly distributed around the central
rays in each bundle.  This means that even when the image is out of
focus and the cone of rays around each central ray do not have their
apex perfectly on the focal plane, the image size does not change.  The
downside of this is, of course, that for all of the central rays to be
perpendicular to the focal plane, the rear element of the lens must be
at least as large as the focal plane diameter.  To get all of the
peripheral rays evenly distributed around the central rays in any bundle
then the rear lens diameter has to be even bigger, at the size of the
focal plane plus the diameter of the lens stop based on the focal length
and f/#.  This leads to very large rear elements on the lens - and
requires a lens mount which is enormously large as well.  So very few,
if any, dSLR lenses are truly telecentric, although some are closer to
telecentric than others, depending on the size of the rear element.

There is an argument which appears to have little, if any, merit that
the sensitivity of digital sensors reduces for light arriving at the
focal plane off of perpendicular.  This would mean that a conventional
lens would produce a reduced response at the edge and corners of the
frame than a telecentric one.  This is one of the main arguments Olympus
have for their 4/3 format - that they can use telecentric lenses with
the small sensor, which would be impossible with larger sensors.  Well,
it turns out that if you look at the Olympus 4/3 lenses they are barely
any more telecentric than their competitors, even on the larger sensors.
Also, if you measure the sensitivity to the angle of incident light that
it is possible to create on a dSLR (where the lens rear element has to
be far enough away from the focal plane to clear the mirror) then it
just isn't significant at all.  I just conducted such a measurement on
the full frame Canon 5D, and the effect is negligible.

So the bottom line is that there is a lot of talk about how important a
telecentric lens is for digital sensors, but the reality is that for
dSLR photography you don't need to worry about it at all.  For example,
I use my old 18mm f/3.5 Olympus OM lens on the Canon 5D quite a lot.
With its 10mm real element (wide open) this lens is almost as far from
being telecentric as it is possible to get.  The vignetting produced is
no different on the Canon, as far as can be seen by eye, from the
vignetting it produced on film - and if telecentricity was important
then that difference would be obvious in this extreme example.  Another
example of how negligible this effect is to look at some of the
vignetting plots from lenses and see how they change at different f/#s.
Strangely, they mainly get better at increasing f/#, yet the principle
ray angle doesn't change, only the cone of the peripheral rays, as f/#
increases.  So the lens creates the light fall off, not the sensors.

Telecentric lenses are hyped for dSLR use - its photographic snake oil.

I agree.  Your statement appeared far more broad than that.  I think a couple
of others read it the same way.

I don't recommend any DX lens on a 35mm camera.  The 12-24 has a reputation of
working and may be the only one where it is practical to try.  In general, it
is a waste of time to put a DX lens on a 35mm camera unless you are trying to
learn something about the lense (like Thom did above).