Moore's law - does it apply to digital SLRs?
Ed Blagden
months. Hence a computer bought 3 years ago for $1000 will have
around one quarter the power of a computer bought for the same price
today. The law has worked pretty well over the last 30 years.
Discussion:
Given that SLR Digital cameras have now reached 10+ MP and 35mm full
frame coverage, and are delivering image quality close to Medium
Format, it's highly unlikely that the market will demand yet more MP
in years to come.
Does this mean that Moore's law will work on pricing rather than
delivering yet more "power" which is not needed? In other words, will
the price of a 10+MP full frame sensor halve every 18 months?
I'm interested in this because right now a top end Digital SLR sells
for $7000+ and a high end film SLR for around $800. The question is
if/when the price of a 10MP digital SLR will drop to $1000 or
thereabouts. For an amateur like me, I would love to play with the
new toys, but shelling out the price of a small car for a digital
camera is crazy and can't possibly be justified (or rationalized, as
my wife would say) in terms of savings on film.
Corrolary to the question is when will 35mm film go the way of vinyl
LPs?
Interested to hear opinions.
Ed
ps This is *not* intended to spark off a "is digital better than
film" war. More about speculating about the future.
Tony Spadaro
size, price, speed and other abilities. Moore's Law is holding but you will
notice that monitors have not exactly kept up with the speed of processors -
and there is good reason to have higher res in monitors. I suspect the 11-14
megapixel full frame SLRs will be as much as anyone needs for a 35mm
replacement - but how about in a package the size of a Rebel Ti with the
frame rate of an EOS 1V complete with booster, and the ability to do
brackets close to instantly, along with an electronic viewfinder etc etc etc
all for 500 bucks? I'm sure it's on the way, but res is not the only factor
in the game.
--
http://chapelhillnoir.com
and partial home of
The Camera-ist's Manifesto
The Links are at
http://home.nc.rr.com/tspadaro/links.html
"Ed Blagden" <ebla...@africaonline.co.ke> wrote in message
news:8e399763.03012...@posting.google.com...
Richard
My understanding of semiconductor fabrication techniques is ropey at best,
but I don't think you'll see the price of full-frame 35mm CCD sensors
dropping as fast as you might expect for, eg. computer memory.
The reason is that the density of "defects" on the silicon wafers which the
chips are made from is roughly constant from wafer to wafer. Secondly the
price of the wafer is roughly constant too. This means that the smaller you
can make the chips, the more you can get from a given wafer, and the higher
the percentage will be free from defects. These two factors taken together
mean that the smaller the chip, the cheaper it will be.
As I understand it, the defect density and the price of the bare wafers
doesn't drop very fast. To push down the price of their products, the makers
of computer chips concentrate heavily on making the transistors the chips
are built from smaller and smaller. A chip with a given function needs a
given number of transistors. So the smaller you make the transistors, the
smaller (and therefore cheaper) the chips will be. There are other reasons
for shrinking the transistors, eg. they run at lower voltage, so run cooler,
and so can be switched faster. Hence multi-gigahertz CPUs.
The problem for digital SLRs is that the size of the sensor is pretty much
fixed by the plate scale of the lens. Unless you change all your lenses
every time a new generation of CCD comes along, you're stuck with a given
size of chip. A chip the size of a 35mm frame (35mm x 24mm?) is absolutely
*massive* compared to run-of-the-mill computer chips. Even the gargantuan
Itanium 2 is only about 20mm x 20mm (and that chip will cost several
thousand dollars each to produce). The Pentium 4 will be even smaller. This
large size of a 35mm CCD sensor means that the defect density on the silicon
wafer really kills the yields, and hence sets a lower limit to the price one
might expect to pay for a digital SLR.
So all in all, I'd be very suprised if we'll see the price of
full-(-ish)-frame digital SLRs drop as fast as computer hardware. For the
point-and-shoot market, you could conceivably re-engineer your lenses to
suit a smaller sensor. It might also be possible to design the sensors to be
fault tolerant, so you'd end up with a full working chip with a few dead
pixels which are massaged away in software (maybe this already happens?).
Like I said, all the above might well be nonsense. If so, I dare say someone
will point it out ;-)
"Ed Blagden" <ebla...@africaonline.co.ke> wrote in message
news:8e399763.03012...@posting.google.com...
Drew
$300 ( wishful thinking, but PIII laptop 512 RAM is about $700 compare the
$3K when it just came out). By that time, I wonder who will dominate the
market.
Capitalism is wonderful.. just a thought ;)
"Ed Blagden" <ebla...@africaonline.co.ke> wrote in message
news:8e399763.03012...@posting.google.com...
Steve Marshall
Pentium 4, for example, is a mind-bogglingly complex beast, with lots of
different subsystems being laid down on the chip, and all having to
communicate with each other. An imaging sensor is a simple (?!) array of
identical elements. My understanding of semiconductor fabrication is pretty
sketchy too, but this seems a significant difference to me.
"Richard" <rgw...@ntlworld.com> wrote in message
news:PI7X9.67$RZ....@newsfep4-win.server.ntli.net...
Richard
"Steve Marshall" <ste...@dialix.com.au> wrote in message
news:3e2d...@news.comindico.com.au...
> Good thoughts, but this may not be comparing apples with apples, exactly.
A
> Pentium 4, for example, is a mind-bogglingly complex beast, with lots of
> different subsystems being laid down on the chip, and all having to
> communicate with each other. An imaging sensor is a simple (?!) array of
> identical elements. My understanding of semiconductor fabrication is
pretty
> sketchy too, but this seems a significant difference to me.
>
Yes, indeed. Furthermore the size of a CCD pixel is quite large (tens of
microns) compared to the size of a transistor in eg. a CPU (hundreds, soon
tens, of nanometres). So CCD fabrication might be more tolerant of defects
in the wafer.
Pschings
>delivering yet more "power" which is not needed? In other words, will
>the price of a 10+MP full frame sensor halve every 18 months?
I don't think so, and here's why: Moore's law has been achieved largely because
of the ability to continuosly reduce the size of chip tracings/transistors,
etc... This will still apply to many of the components in a digital SLR, but
not in the sensor. As long as the market demands 24x36mm sensors the
manufacturers will find it much more difficult to increase chip yields from the
wafers, at least for these components. So I suspect that unless (until?)
radically different manufacturing methods are invented to increase yields for
large area components, the prices of digital SLRs won't drop as fast as Moore's
law would suggest. They will drop as demand and production increases, but not
as dramtatically as other elecronic devices.
Of course, I'm probably wrong, but that's my take on it.
Paul
Don Stauffer
for fabrication of digital logic circuits.
A CCD is actually an analog IC, in spite of it being in a digital
camera. Digitization is done AFTER the CCD is read out.
The critical portion of the CCD for our purposes are the image sensors
themselves. They must collect a certain minimum number of photons
during the exposure, and hence are sensitive to area. Too small and
they get noisy.
So, while improvements in semiconductor fabrication may help some of the
ancillary circuits on the chip, don't look for the density of the
photosensors to increase that much.
A general lowering of semiconductor processing costs can, however, make
it more economical to make larger chips, and this would certainly be a
boon to digital cameras.
--
Don Stauffer in Minnesota
stau...@usfamily.net
webpage- http://www.usfamily.net/web/stauffer
Gregory L. Hansen
Richard <rgw...@ntlworld.com> wrote:
>
>My understanding of semiconductor fabrication techniques is ropey at best,
>but I don't think you'll see the price of full-frame 35mm CCD sensors
>dropping as fast as you might expect for, eg. computer memory.
Moore's first law is that the density of transistors will double every
eighteen months. Moore's second law is that the cost of a fab is
proportional to the density of transistors it can produce. So far, I
think, both laws are holding up pretty well.
--
"A nice adaptation of conditions will make almost any hypothesis agree
with the phenomena. This will please the imagination but does not advance
our knowledge." -- J. Black, 1803.
Chris Quayle
>
<snipped>
> This
> large size of a 35mm CCD sensor means that the defect density on the silicon
> wafer really kills the yields, and hence sets a lower limit to the price one
> might expect to pay for a digital SLR.
>
But you can get around the problem of defects by using (for example)
software to keep a map / lookup table of defects, which are then filled
in by interpolating the surrounding pixels. Wouldn't be surprised if
this isn't being done already, but I doubt if the manufacturers will
admit it :-(.
Using techniques like that means that sensors can be graded according to
the market the camera is aimed at and it's price range...
Chris
Paul Farrar
Bingo.
And a lot of the noise is due to statistical mechanics of the
devices; so getting the desired S/N ratio requires a certain
size. (Unless you cool the chip, which astronomers, at least,
often do.)
Paul
ThomasH
>
> Moore's law does not strictly apply to CCD chips per se. It is primarily
> for fabrication of digital logic circuits.
>
> A CCD is actually an analog IC, in spite of it being in a digital
> camera. Digitization is done AFTER the CCD is read out.
I think that this is the important point here and that's why the
resolution of opto-semiconductor was never as rapidly growing as
was the size and speed of transistors in CPU's. But now with the
billions of sales in digital cameras of all sorts the CDD and
CMOS might be on a fast lane. Personally I believe that our
mammal and birds eyes are made of a better material and if my
memory serves me correctly, Kodak experiments already with
prototypes of organic sensors.
Thomas
Robert Whitehouse
way that we approach photography and cameras. Todays crop of digital cameras
are essentially digital versions of film cameras.
Slowly (because the markets can't adapt that quickly) we will see our use of
cameras evolve in ways that we cannot even yet imagine. Witness the PC
revolution - who would have predicted that a primary use of PCs would be
"surfing the web" - this would have had no meaning in 1980 when PCs were at
the same stage that digital cameras are today.
Sensors will undoubtedly improve but I see cameras getting smarter and the
big advances coming through new "applications".
Perhaps one area that we might see is in automated contrast control - why
do todays sensors suffer the same draw-back as film i.e. that the sensor
has the same sensitvity (speed) across its whole area ? - why not have a
"smart" sensor that can do automatic contrast compression where required (or
at user request) ?
This is just one example - perhaps you guys can think of more ? (it might
give the camera makers some food for thought - if they are listening !)
BW
"Ed Blagden" <ebla...@africaonline.co.ke> wrote in message
news:8e399763.03012...@posting.google.com...
Dallas
soon. We don't need bigger sensors, because the one in the EOS1Ds and the
soon to be announced Nikon equivalent will do everything that is required of
it (compared to 35mm that is).
Maybe all they really need to do now is address issues of writing speed and
memory capacity. I'm willing to bet that the next R&D area they will
concentrate on will be one of media. Expect CF cards to be replaced with
something bigger and better, thus necessitating the need to upgrade your
camera to take advantage of that innovation. This appears to be the Sony way
of doing things, no reason to think that Canon & Nikon won't do the same.
I think the question every photographer looking to go digital should be
asking about the D-SLR they are considering is this:
Is this as good as it gets?
If yes, then great. If no, then I guess you should be playing the waiting
game.
"Ed Blagden" <ebla...@africaonline.co.ke> wrote in message
news:8e399763.03012...@posting.google.com...
Richard Cochran
> Moore's law does not strictly apply to CCD chips per se. It is primarily
> for fabrication of digital logic circuits.
>
> A CCD is actually an analog IC, in spite of it being in a digital
> camera. Digitization is done AFTER the CCD is read out.
>
> The critical portion of the CCD for our purposes are the image sensors
> themselves. They must collect a certain minimum number of photons
> during the exposure, and hence are sensitive to area. Too small and
> they get noisy.
>
> So, while improvements in semiconductor fabrication may help some of the
> ancillary circuits on the chip, don't look for the density of the
> photosensors to increase that much.
Correct. Furthermore, there's the issue that it doesn't make
sense to pack sensors in so tightly that the lens isn't able to
guide the photons correctly to the right pixel! At small to middle
apertures, many 35mm lenses are already approaching the diffraction
limit to resolution, caused by the aperture and the wavelength of
light. Most 35mm lenses aren't diffraction limited wide open,
but it would still be expensive and difficult to improve the
correction very much beyond what they currently deliver.
IC fabricators use lenses to create the ICs, so they're limited by
the same laws of physics that photographers are. But they can
"cheat", by using wavelengths shorter than visible light. They
long ago had to quit using visible light and move into ultraviolet
and X-rays, and they keep moving to still shorter wavelengths.
Unfortunately, photographers are pretty much stuck with the
wavelengths present in visible light. So the only way to
improve usable resolution beyond what fine-grained film is
already delivering is to move to a larger sized image sensor.
If you instead move to a smaller sized sensor, you'll decrease
usable resolution, regardless of pixel count. If you could
make 2mmx3mm CCD with 10 megapixels worth of sensors, no lens
could produce a sharp 10 megapixel image on it, at least not
using visible light.
--Rich
Rudy Garcia
ebla...@africaonline.co.ke (Ed Blagden) wrote:
> Moore's law says that computing power per $ doubles roughly every 18
> months. Hence a computer bought 3 years ago for $1000 will have
> around one quarter the power of a computer bought for the same price
> today. The law has worked pretty well over the last 30 years.
>
That isn't exactly what Moore's law says. It states that transistor
count doubles every 18 months (after it was revised from the original
which doubled transistor count every 12 months). Moore's law makes no
prediction about PC prices. It is also not a "law" per se, but an
empirical observation. Although it is amazing that it has worked all
this time, It is questionable that it can continue unabated.
> Discussion:
>
> Given that SLR Digital cameras have now reached 10+ MP and 35mm full
> frame coverage, and are delivering image quality close to Medium
> Format, it's highly unlikely that the market will demand yet more MP
> in years to come.
>
> Does this mean that Moore's law will work on pricing rather than
> delivering yet more "power" which is not needed? In other words, will
> the price of a 10+MP full frame sensor halve every 18 months?
Chip volumes is what drives the price and the chip manufacturing
learning curve. For example, as chip manufacturing moves to the next
process node (130 nanometers), a full mask set for a complex CMOS device
will run something over $1 million bucks. There better be a large
volume to justify the design development costs and the tooling to build
the chip.
> I'm interested in this because right now a top end Digital SLR sells
> for $7000+ and a high end film SLR for around $800. The question is
> if/when the price of a 10MP digital SLR will drop to $1000 or
> thereabouts. For an amateur like me, I would love to play with the
> new toys, but shelling out the price of a small car for a digital
> camera is crazy and can't possibly be justified (or rationalized, as
> my wife would say) in terms of savings on film.
>
I dare think that even on today's digital cameras the total
semiconductor content is but a fraction of the total cost. Most of the
cost is in the machined parts, optical components and distribution
channel costs.
> Corrolary to the question is when will 35mm film go the way of vinyl
> LPs?
>
Not anytime soon.
> Interested to hear opinions.
>
You got them.
> Ed
>
> ps This is *not* intended to spark off a "is digital better than
> film" war. More about speculating about the future.
--
Rudy Garcia
Rudy Garcia
"Steve Marshall" <ste...@dialix.com.au> wrote:
> Good thoughts, but this may not be comparing apples with apples, exactly. A
> Pentium 4, for example, is a mind-bogglingly complex beast, with lots of
> different subsystems being laid down on the chip, and all having to
> communicate with each other. An imaging sensor is a simple (?!) array of
> identical elements. My understanding of semiconductor fabrication is pretty
> sketchy too, but this seems a significant difference to me.
What you say about complexity is very true, but random defects don't
care about complexity when picking a place to land on the die.
What are important is: the defect density, the distribution of defect
size and the critical die area. All of the above are on a per layer
basis.
An image sensor is simple, but it is also a pretty large area.
--
Rudy Garcia
Dennis O'Connor
The pro journalists and advertising/catalog photography will all go totally
digital - without exception... They can file their pictures instantly,
color can be corrected at the click of a mouse, ad copy can be added at the
click of a mouse, etc... Time is money and money rules...
A significant portion of upscale amateur photography will go digital because
WE can afford fast PC's, web sites, color printers, ink, and D60's that
become obsolete before they are paid for, etc... Digital is the moment, it
is what is happening, it is WHAT is being presented in the magazines (the
major opinion maker)
The fine art portion of photography will remain silver based indefinitely,
and fine art prints will become even more expensive and collectible, etc..
Does that mean that no digital photographers will make money in the fine art
or wall hanging world? Nope... It just means that the top end remains
silver, just like the top end of fine art painting remains brush and liquid
paints, even in a computer world...
The lower socioeconomic folks of the world... Ahhh, the barely more than
poor... They are the manna from heaven for film photography... You see, the
hard working guy at the gas station and his wife who is waiting tables at
the local saloon, they have birthdays and weddings like everyone else...
They want pictures... They do not have a fast PC, or an ISP, or a color
printer, and they can't even fantasize about having a personal web site...
They are the bread and butter for film makers, because they will go to the
local supermart and grab a bubble encased plastic or paper camera off the
wall, already loaded with film, flash and batteries, and take pictures of
the events in their lives...Then they drop them back at the supermart for
developing and prints... Prints they can share with their friends and family
who also cannot afford the fast PC, etc... That is not going to change in
the near future...
Denny
"Rudy Garcia" <ru...@jps.net.invalid> wrote in message
news:rudyg-739707....@nnrp04.earthlink.net...
Bill Hilton
>I have said this before, but repetition is good for the soul..
>
>The fine art portion of photography will remain silver based indefinitely,
>and fine art prints will become even more expensive and collectible, etc
Hmm ... I know of at least a dozen "fine art photographers" in my area
(southwest USA) who have switched gears the past few years. Top name
photographers like Tom Till, David Muench, Jack Dykinga, Linde Waidhofer ...
even guys like Galen Rowell and Frans Lanting in California. The local color
boys used to shoot medium format or large format (4x5" or even 8x10") film,
usually Velvia, and make Ilfochrome prints (formerly Cibachromes) using
contrast masks. These prints were gorgeous and sold for hundreds of dollars,
even thousands for large prints in limited editions.
Almost every one of these guys has switched from Ilfochromes to a digital flow
the past 3-4 years, with their film scanned on a drum scanner, tonal and color
adjustments done in Photoshop, and prints done on Fuji Crystal Archive paper
with a LightJet laser printer. The prints done this way offer several
advantages compared to Ilfochromes -- prints are cheaper by almost half,
sharper, you have more control over contrast and color at the Photoshop step
and it's more repeatable. It's also much more enviromentally friendly, and you
don't have to send your original in every time you need a print, just send a
copy of the file (which is usually kept archived at the printer anyway). Also
the estimated print life is about double that of the Ilfochromes, significant
for the fine art market.
Several labs report their Ilfochrome print business is off by as much as 75%
and some excellent labs have stopped doing Ilfochromes altogether. Now you can
buy a $3,000 or $5,000 Epson 7600 or 9600 (24" or 44" wide carriages) that
outputs prints for all practical purposes as good as those from a $200,000
LightJet and that last even longer, with about 3x the estimated print life of
Ilfochromes. Pretty incredible.
I attended a digital printing workshop last week at Calypso Labs in Santa
Clara, taught by the "Ansel Adams of digital printing" and by a pro
photographer who is an expert with dye transfer printing. They showed prints
of the same image printed with a Type R process (not that great), the
Ilfochrome process (nice if you are skilled enough to do the delicate contrast
masking, too contrasty otherwise), printed as a dye transfer (beautiful prints
but it takes a full day to do one, IF you're an expert, and with a bit less
gamut than the best digital processes), printed on the LightJet and printed on
the Epson 9600. When shown side-by-side the Epson and LightJet prints were
clearly the equal of the Ilfochromes, better if the contrast masking wasn't
done 'just so'.
Black-white photographers are moving to digital printing more slowly but look
at what guys like Dan Burkholder are doing and you'll notice a trend there too.
Personally I think the contrast masked Ilfochromes of Christopher Burkett are
probably the finest prints I've ever seen (he shoots 8x10" Velvia and does his
own darkroom work, shooting 2 months a year and printing 10 months!) and I'd
love to have one as a goal to strive for. And I've often visited a fine art
photo gallery in Santa Fe the past few years and watched as the better
black-white artists' work has sky-rocketed ... when I was there in December
they had several Ansel Adams prints ranging from $20,000 to $75,000 and the
Henri Cartier-Bresson prints that sold for $3,000 two years ago were $5,500
this year. So there will be a place for the finest quality works, for those
collectors who can afford them.
But the trend to digital for new fine art work seems to be pretty definitive,
especially for color.
Bill
Bill Tuthill
>
> Does this mean that Moore's law will work on pricing rather than
> delivering yet more "power" which is not needed? In other words,
> will the price of a 10+MP full frame sensor halve every 18 months?
No.
Moore's law is based on shrinking chip and die size. The future is
probably full-frame CMOS sensors, or possibly half-frame CCD sensors
as well, and neither can be shrunk due to optical limitations.
Yields will be low at first. We have seen Canon's supply problems.
Yields might increase over time, but not quickly. I would not expect
prices to drop 50% every 18 months. Prices will decrease only when
manufacturing yields improve.
Gordon Moat
phone to complete the kit on location. Quite a few journalists are also going
video, and just using the stills if needed, or just having saleable footage.
The catalogue world is high enough volume to justify the expensive digital
backs, though many of those are leased. The catalogue world also has some
flexibility on meeting deadlines, at least enough to use film, but the profits
are from volume. the lease plans can allow for technology upgrades, though I
have seen CDs full of 10 MB or smaller files mostly, so not really pushing
technology that much.
Magazines and advertising are a different story. There is a mix currently, with
lots of film still being scanned. If development stopped on scanners, then
direct digital might be better. With on location shoots, one can satisfy ADs
with Polaroids, postage stamp sized LCDs, digital to a laptop screen, or video.
Some architecture shooting is still done with 4" by 5" cameras, and the film is
then drum scanned.
The video I thought was the most interesting proofing set-up, and I have seen
two different set-ups with this. Each photographer had a miniDV camcorder going
to a small external monitor. The AD was allowed to control the video camcorder
to review any of the session. The difference between the two photographers was
that one clamped the camcorder onto the same tripod used for the shoot, giving
almost the same field of view.
One nice side effect of digital SLRs is that the small files have made scanned
35 mm film more acceptable, especially in situations that used to be strictly
medium format shooting. Better and more affordable film scanners have been
replacing drum scans for quite a bit of commercial work.
The oddest aspect of current digital files, is the usage for various forms of
advertising. With many companies going in house on graphic design, or hiring
poorly qualified temporary workers, the has been a proliferation of errors, and
a drop in the quality of some printed materials, especially compared to just
five years ago. I have seen large corporations take thumbnail sized digital
images, and blow them up 20+ times for printing on advertising banners. I have
also seen many magazine ads with poorly upsized images still showing edge
aliasing, or artefacts, the worst example from Toyo for their new Large Format
camera advertisements in some magazines.
It is truly anything goes at this point in time. Whatever you can use to meet
the deadlines and still have some profits. Lights and computer gear can often
cost far more than the camera bodies. Lenses are still expensive, but computer
resizing, cropping, and stitching are eliminating the need to have lots of
lenses.
Companies that are more concerned with deadlines than quality will take almost
anything for advertising uses. The ones that do bias quality will get the least
complaints, though perhaps not the most work. Modern printing technology
evolves slowly, and there are many ways to match it.
Okay, a funny story. I had one client about a year ago that was adamant about
wanting digital images on CD-R for the advertising images I was shooting. Along
with the bags of gear I brought, I had rigged one of the camera bodies with a
high tech looking bit, that actually did nothing. I told the client rep that
was the digital camera, and the film camera was just to back up and some low
light shots. They already knew that colour correction, and pre press
preparation, would take me a couple days, so they were not expecting images
when I left the location. I scanned the film, did all the corrections for
printing, and delivered a CD-R on time. They were very happy with the quality
of the images. Anyway, one point is that scanned film is still digital.
Your comment about one time use cameras is right on the mark. This is the
highest profit and volume area for Kodak and Fuji (I think for AGFA as well).
These will likely continue to be a fixture for many years, until the
manufacturers come out with disposable digital.
Digital was the moment when 2 MP Kodak digital SLRs were all the rage, at over
$10000. Cheap digital point and shoots are the current rage, though often just
for screen display, internet, or a weird toy. There is a generation of
photographers taking pictures at arms length, many with cameras that get
smaller every year.
Despite that this is an equipment news group, the skill of the individual
taking the photos often dictates what gets published. Like I stated before,
anything goes. I have seen pinhole shots in galleries, blurred and underexposed
shots for major fashion shoots, a book published from images taken with
disposable cameras, and awful amateur crap with multi thousand dollars gear.
Just like buying a Harley will not make one a biker, buying a digital camera
will not make one a PhotoShop expert. Though if you want to join one group, or
the other, practice, practice, practice.
Ciao!
Gordon Moat
Alliance Graphique Studio
<http://www.allgstudio.com>
Woody W.
You might not have noticed, but Fuji just introduced a new ccd that extends
dynamic range by having two discrete sensors per photosite, one large and
high-sensitivity, one small with low-sensitivity.
- Woody -
"Robert Whitehouse" <bob.whi...@btinternet.com> wrote in message
news:b0ka29$5n5$1...@knossos.btinternet.com...
<<clip>>