«

Don Wiss <donwiss@no_spam.com> wrote:
> On Sat, 15 Dec 2007, The Good Doctor <> wrote:

>>Neil Ellwood <> wrote:
>>
>>>Don Wiss wrote:
>>>
>>>> I'd like to see someone create a Google mash-up that tells me when
>>>> the sun is in the direction I want. On the map you would put two
>>>> points. One for where you would be standing. The second would be off
>>>> in the direction you want to be looking. Then you would input a
>>>> date. The program would return the time on that date that the sun
>>>> would be directly behind you.

You don't need the date. The time at which the sun will be directly
behind you is always the same, all that changes is the summer time
adjustment and the height in the sky. Otherwise sundials wouldn't
work.

>>>Use a compass and a watch.
>>
>>Learn to use your watch as a compass, and a watch is all you need.
>>It is all I need, and it gets plenty of use.

> I'm sorry but I don't follow. How can that tell me at what time I should go
> visit a site to take a picture with the sun exactly in the direction I want
> it to be?

The earth spins 360 degrees in 24 hours. An analogue watch face is
conveniently graduated in units of 1/60th of a circle. It also tells
you the time now. So if you can see the sun now, it's easy to use the
watch face to show you where it will be at any later or earlier time
in the day. Or when it will be in any particular required direction.

Since the earth revolves in 24 hours and the watch hour hand in 12 the
hour hand moves twice as fast as the sun. At 6am and 6pm local solar
time the sun is due east or west, although sometimes, depending on
latitude, it will be below the horizon. You could make yourself a sun
position protractor graduated in time you can just drop over your
local map.

I was taught this kind of thing when I was taught map reading in
school. But then I'm so old that when I was at school they still
believed in educating children

--
Chris Malcolm DoD #205
IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK
[http://www.dai.ed.ac.uk/homes/cam/]

[A complimentary Cc of this posting was sent to
Chris Malcolm
<>], who wrote in article <>:
> >>>> date. The program would return the time on that date that the sun
> >>>> would be directly behind you.
>
> You don't need the date. The time at which the sun will be directly
> behind you is always the same, all that changes is the summer time
> adjustment and the height in the sky. Otherwise sundials wouldn't
> work.

Did you see a curve somewhat like "figure 8" drawn near equator on
some globes? My conjecture is that it shows the position of sun at
(local) midday relative to earth.

If so, the width of this curve is how much your conjecture errs.

(Hint: the orbit of Earth about Sun is not circular; thus in some
parts the orbital speed is quickier, in some slower. Effectively, the
1/365.25day [+-1] [sp? ;-] movement of Sun about Earth is not with
constant angular velocity).

Hope this helps,
Ilya

acl wrote:
> On Dec 15, 6:19 am, Ilya Zakharevich <nospam-ab...@ilyaz.org> wrote:
>> [A complimentary Cc of this posting was sent to
>> Charles
>> <ckr...@SPAMTRAP.west.net>], who wrote in article <8ma6m3d6b53dj9kmtgso80ke0hsgiao...@4ax.com>:
>>
>>> Not quite what you are asking for, but you may get some use from this
>>> page.
>>> http://aa.usno.navy.mil/data/docs/AltAz.php
>> All I get is
>>
>> Invalid year. Please enter a value from 1700 to 2100
>>
>> (I was trying 3007). Anybody hear knowing: with the current precision
>> of data about sun/earth/moon system, for how long one can do reliable
>> prediction (say, with 15' error)?
>>
>> [I know that for Solar system as a whole, chaos appears in about 1e9
>> years increments (at least, this was state-of-art of 1990); but
>> even without chaos, errors in measurements would matter at some
>> moment...]
>
> Yes, that's funny isn't it, I know about Sussman and Wisdom's work,
> but have no clue about the accuracy to which this stuff is known in
> 100 years or so. Strange world.

Every 1 or 2 years, there's a "leap second". This is to correct
for the varying length of the day. That depends on how fast
the earth rotates about its axis. The other planets and the moon
"don't care" how fast or slow the earth is rotating;
they just follow their orbits around the sun (or earth for the moon).

The figure-8, or analemma, is explained in detail at
< http://www.analemma.com/ > . There's a free program there,
SunGraph, which I used to have istalled before I reinstalled Windows.
It's pretty good.

David Bernier

On Dec 18, 2:07 am, David Bernier <david...@videotron.ca> wrote:
> acl wrote:
> > On Dec 15, 6:19 am, Ilya Zakharevich <nospam-ab...@ilyaz.org> wrote:
> >> [A complimentary Cc of this posting was sent to
> >> Charles
> >> <ckr...@SPAMTRAP.west.net>], who wrote in article <8ma6m3d6b53dj9kmtgso80ke0hsgiao...@4ax.com>:
>
> >>> Not quite what you are asking for, but you may get some use from this
> >>> page.
> >>>http://aa.usno.navy.mil/data/docs/AltAz.php
> >> All I get is
>
> >> Invalid year. Please enter a value from 1700 to 2100
>
> >> (I was trying 3007). Anybody hear knowing: with the current precision
> >> of data about sun/earth/moon system, for how long one can do reliable
> >> prediction (say, with 15' error)?
>
> >> [I know that for Solar system as a whole, chaos appears in about 1e9
> >> years increments (at least, this was state-of-art of 1990); but
> >> even without chaos, errors in measurements would matter at some
> >> moment...]
>
> > Yes, that's funny isn't it, I know about Sussman and Wisdom's work,
> > but have no clue about the accuracy to which this stuff is known in
> > 100 years or so. Strange world.
>
> Every 1 or 2 years, there's a "leap second". This is to correct
> for the varying length of the day. That depends on how fast
> the earth rotates about its axis. The other planets and the moon
> "don't care" how fast or slow the earth is rotating;
> they just follow their orbits around the sun (or earth for the moon).
>
> The figure-8, or analemma, is explained in detail at
> <http://www.analemma.com/> . There's a free program there,
> SunGraph, which I used to have istalled before I reinstalled Windows.
> It's pretty good.

Thanks. What I had in mind is something like "to what accuracy can I
predict the position of the planets in 100 years?". ie, how accurately
are their positions determined? Anyway, I was just expressing suprise
at my lack of knowledge! I'm sure these things can be found out (eg by
asking astronomers).

Ilya Zakharevich <nospam-> wrote:
> [A complimentary Cc of this posting was sent to
> Chris Malcolm
> <>], who wrote in article <>:
>> >>>> date. The program would return the time on that date that the sun
>> >>>> would be directly behind you.
>>
>> You don't need the date. The time at which the sun will be directly
>> behind you is always the same, all that changes is the summer time
>> adjustment and the height in the sky. Otherwise sundials wouldn't
>> work.

> Did you see a curve somewhat like "figure 8" drawn near equator on
> some globes? My conjecture is that it shows the position of sun at
> (local) midday relative to earth.

> If so, the width of this curve is how much your conjecture errs.

> (Hint: the orbit of Earth about Sun is not circular; thus in some
> parts the orbital speed is quickier, in some slower. Effectively, the
> 1/365.25day [+-1] [sp? ;-] movement of Sun about Earth is not with
> constant angular velocity).

THat's true, and important if you're navigating with a sextant or
doing astronomy, but for the purposes of working out when the sun will
be at your back when you're photographing the front of a building it's
irrelevant.

--
Chris Malcolm DoD #205
IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK
[http://www.dai.ed.ac.uk/homes/cam/]

acl wrote:
> On Dec 16, 12:18 am, Ilya Zakharevich <nospam-ab...@ilyaz.org> wrote:
>> [A complimentary Cc of this posting was NOT [per weedlist] sent to
>> acl
>> <achilleaslazari...@yahoo.co.uk>], who wrote in article <021f584f-f8ad-4fab-b21e-72c87261b...@d4g2000prg.googlegroups.com>:
>>
>>>> (I was trying 3007). Anybody hear knowing: with the current precision
>>>> of data about sun/earth/moon system, for how long one can do reliable
>>>> prediction (say, with 15' error)?
>>>> [I know that for Solar system as a whole, chaos appears in about 1e9
>>>> years increments (at least, this was state-of-art of 1990); but
>>>> even without chaos, errors in measurements would matter at some
>>>> moment...]
>>> Yes, that's funny isn't it, I know about Sussman and Wisdom's work,
>>> but have no clue about the accuracy to which this stuff is known in
>>> 100 years or so. Strange world.
>> Aha, thanks for the reference (I presume you mean "Chaotic Evolution
>> of the Solar System" of 1992).
>
> Yes, although I haven't really read it (far too hairy for my taste,
> and I'm not an efficient paper-reader anyway). I just remembered
> someone telling me the conclusions a few years ago. I misremembered,
> too
>
>> What I remembered was slightly earlier
>> numerical experiments; they claimed that all the planets except Pluto
>> were not showing chaotic behaviour for 1e9 scale, and only the
>> trajectory of Pluto was chaotic; this was about '89 [I looked it up,
>> and I suspect it was from "Numerical evidence that the motion of Pluto
>> is chaotic" of '88; they say 2e7 years, not 1e9 I remembered]. I
>> forgot about the newer result; they get much higher dynamic entropy,
>> 1/(4e6 years)...
>>
>> Now to read the whole article [hmm, Science copy not accessible
>> through my usual channels...]. Fortunately, we have not only
>> academia, but also the armed forces:
>> http://handle.dtic.mil/100.2/ADA260055
>
> Has there not been more recent work on this? It's sort of interesting
> (the conclusions are, anyway).

Maybe you want to know accuracy for time on the atomic clock time
scale. If so, the accuracy will be much better than for
non-atomic time (GMT, UTC)...

All this depends on whether one is using
"GMT"-type time (solar), or TAI (international
attomic time). The positions in the future
are much better known for TAI, but
UTC ( which is "GMT"-type time) is tweaked
atomic time so the sun is near the meridian
at noon local time. In the long run,
"GMT"-type time is the best for humans ...

To get the position of the moon at 12h00 GMT
in London on Jan 1, 1000 BC, we need to know
about the length of the day since 1000 BC.
This is because GMT is based on the length of
a solar day (not atomic time). The length of
a solar day varies by a minute amount (hence,
the leap seconds).

The relevant quantity is Delta T, or the
earth's clock error. Morrison and Stephenson
have a paper which is linked to from
< http://en.wikipedia.org/wiki/Delta_T >
It's the one in the Journal for the History of Astronomy.

They estimate Delta T in seconds, and their uncertainty
sigma in the value of Delta T in Table I.

For -1000, (or 1001 B.C. I suppose), their uncertainty
is +/- 640 seconds. Relative to the stars, the moon moves
by about 6 arc-minutes in that time, or 0.2 of the apparent
lunar diameter. So for Jan 1, 1000 BC at 12h00 GMT
in London, we would have had an uncertainty of
+/- 6 arc-minutes in the moon's position in the sky.

That was 3000 years ago. So for 12h00 GMT in
5000 AD, a similar uncertainty of +/- 6 arc-minutes
in the moon's position is a guess.

David Bernier

On Dec 18, 6:32 am, David Bernier <david...@videotron.ca> wrote:


>
> Maybe you want to know accuracy for time on the atomic clock time
> scale. If so, the accuracy will be much better than for
> non-atomic time (GMT, UTC)...

Yes, that's what I was asking. But the "backwards" approach is also
interesting. I took a quick look at Morrison-Stephenson, I'll read it
carefully tomorrow.

Thanks!

[A complimentary Cc of this posting was sent to
Chris Malcolm
<>], who wrote in article <>:
> > (Hint: the orbit of Earth about Sun is not circular; thus in some
> > parts the orbital speed is quickier, in some slower. Effectively, the
> > 1/365.25day [+-1] [sp? ;-] movement of Sun about Earth is not with
> > constant angular velocity).
>
> THat's true, and important if you're navigating with a sextant or
> doing astronomy, but for the purposes of working out when the sun will
> be at your back when you're photographing the front of a building it's
> irrelevant.

The width of analemma (thanks for the word, BTW!) is 8 degrees. So if
30min error is negligible, then one does not need to know anything -
just come in the morning, and some time soon the sun will be at your
back.

Yours,
Ilya

Ilya Zakharevich <nospam-> wrote:
> [A complimentary Cc of this posting was sent to
> Chris Malcolm
> <>], who wrote in article <>:
>> > (Hint: the orbit of Earth about Sun is not circular; thus in some
>> > parts the orbital speed is quickier, in some slower. Effectively, the
>> > 1/365.25day [+-1] [sp? ;-] movement of Sun about Earth is not with
>> > constant angular velocity).
>>
>> THat's true, and important if you're navigating with a sextant or
>> doing astronomy, but for the purposes of working out when the sun will
>> be at your back when you're photographing the front of a building it's
>> irrelevant.

> The width of analemma (thanks for the word, BTW!) is 8 degrees. So if
> 30min error is negligible, then one does not need to know anything -
> just come in the morning, and some time soon the sun will be at your
> back.

Thirty minutes sounds like a lot, but the point is that the real
estate photographer's general rule of "sun at your back" has a lot
more more than eight degrees of tolerance in sun position in it. And
if you do happen to be fussier than that, such as wanting the sun to be
just grazing a wall to highlight texture, then a simple recent
sighting of the sun's noon position will enable you to cut down that
error by a factor of at least ten which should be plenty.

What interests me about this thread is how much cheap computing power
and web access has taken over from simple common sense. It wouldn't
surprise me these days to hear the following conversation from two
tourists standing in a street in the sun:

"Which direction is north?"

"Wait I minute while I get Google Earth up on this thing and I'll
tell you."



--
Chris Malcolm DoD #205
IPAB, Informatics, JCMB, King's Buildings, Edinburgh, EH9 3JZ, UK
[http://www.dai.ed.ac.uk/homes/cam/]

[A complimentary Cc of this posting was sent to
Chris Malcolm
<>], who wrote in article <>:
> > The width of analemma (thanks for the word, BTW!) is 8 degrees. So if
> > 30min error is negligible, then one does not need to know anything -
> > just come in the morning, and some time soon the sun will be at your
> > back.
>
> Thirty minutes sounds like a lot, but the point is that the real
> estate photographer's general rule of "sun at your back" has a lot
> more more than eight degrees of tolerance in sun position in it.

Then the same argument applies: if 8degrees error is negligible, it is
not a problem to come at any time. Just come before the noon if the
shot direction is to the west of north, and in afternoon if it is to
the east of north - and most of the bases are covered.

> What interests me about this thread is how much cheap computing power
> and web access has taken over from simple common sense.

What interests me about this thread is how much error the "simple
common sense" would bring. [I saw the analemma drawn on maps; I
puzzled about it, then deduced what it should mean; but I never paid
attention on HOW WIDE it is! Only yesterday, looking on Wikipedia, I
found that error might be up to half an hour; my guts feeling would
bring it to an order of magnitude of a minute.]

Hope this helps,
Ilya