# What's the time?

I like to save time quite literally, by wearing three watches on my wrist... Problem is they each give a different time. One watch is x minutes behind the actual time. One watch is x minutes ahead of the actual time. The last watch shows the actual time.

Problem is, I can't tell which watch has the correct time...

From the time displayed on each watch, determine the actual time. If it is not possible to determine the time, print "Look at the sun".

Input: Three readings, separated by single space characters: H1:M1 H2:M2 H3:M3 In each reading H1,H2,H3 represent the hours displayed (0 < H1,H2,H3 < 13), and M1,M2,M3 represent the minutes displayed (0 <= M1,M2,M3 < 60). If the number of minutes is less than 10, a leading 0 is prepended in the input. Similarly, is the number of hours is less than 10, a leading 0 is prepended in the input.

Output: The correct time is HH:MM where HH:MM is the correct time. If no correct time can be determined, it displays Look at the sun.

Input 1: 05:00 12:00 10:00

Output 1: The correct time is 05:00

Input 2: 11:59 12:30 01:01

Output 2: The correct time is 12:30

Input 3: 12:00 04:00 08:00

Output 3: Look at the sun

Shortest code wins... No special penalties apply. Also, bear in mind that we're dealing with a 12-hour clock... I don't care about AM or PM... Imagine we're dealing with analogue watches...

• By correct time, you mean the current local time ? – Optimizer Jun 3 '15 at 6:21
• @sanchises Because they're 4 hours apart. 12:00 is exactly between 8:00 and 4:00 too, you know... – Siguza Jun 3 '15 at 7:40
• You say the output is of the form The correct time is HH:MM, without a full stop, but then proceed to include a full stop in the first two examples. Which version is correct? – Sp3000 Jun 3 '15 at 15:49
• Is it allowed to to read the times as command line arguments? – Dennis Jun 3 '15 at 16:38
• As in whattimeisit 07:21 08:39 08:00 ? Sure! – WallyWest Jun 3 '15 at 23:43

# CJam, 86837775 71 bytes

"The correct time is "ea{aea+':f/60fb(f-:+720%!},{];"Look at the sun"}*


Thanks to @jimmy23013 for golfing off 6 bytes from my code.

Try it online in the CJam interpreter.

### Test cases

$cjam time.cjam 05:00 12:00 10:00; echo The correct time is 05:00$ cjam time.cjam 11:59 12:30 01:01; echo
The correct time is 12:30
$cjam time.cjam 12:00 04:00 08:00; echo Look at the sun  ### How it works "The correct time is " e# Push that string. ea e# Push the array of command-line arguments. { e# Filter; for each time T: a e# Wrap T in an array. ea+ e# Concatenate with the array of all times. ':f/ e# Split each time at the colon. 60fb e# Consider it a base 60 integer. (f- e# Shift out the converted T and subtract it from the remaining times. :+ e# Add them. 720%! e# Push 1 is the sum is 0 modulo 720 and 0 if not. }, e# Keep the time iff a 1 has been pushed. e# The array on the stack now contains one or three times. { e# Reduce; for each time but the first: ]; e# Discard the entire stack. "Look at the sun" e# Push that string. }* e#  • qS/_':f/60fb_f{f-:+720%!,}\"The correct time is "f\2/.e&$("Look at the sun"@?. b implies :i. – jimmy23013 Jun 3 '15 at 7:17
• "The correct time is "lS/_':f/60fb:T.{Tf-:+720%{}@?}{];"Look at the sun"}*. – jimmy23013 Jun 3 '15 at 8:13
• You missed the case when x=0 – Lakshay Garg Jun 3 '15 at 19:57
• @LakshayGarg: The question says that they each give a different time. That would be false if x = 0. – Dennis Jun 3 '15 at 20:07
• @Dennis You may be able to reduce it further, the full stop at the end isn't required... – WallyWest Jun 3 '15 at 23:46

# JavaScript (ES6), 164 168 172

For each reading, calc the distance from the other two. The one that have the same distance is what you need. If there is more than one, then you can't tell.

F=s=>(s=s.split(' '))
.map(x=>([h,m]=x.split(':'),+m+(h%12)*60))
.map((v,i,z)=>(v+v-z[++i%3]-z[++i%3])%720||'The correct time is '+s[k+=i],k=-2)
[k]||'Look at the sun'

// TEST

out=x=>O.innerHTML += x+'\n';

['12:30 01:01 11:59', '11:50 11:55 12:00', '05:00 12:00 10:00', '12:10 04:10 08:10', '07:00 06:00 08:00']
.forEach(x => out(x + ' -> ' + F(x)))

go=_=>out(I.value + ' -> ' + F(I.value))
<pre id=O></pre>
<input id=I><button onclick="go()">Test</button>

• Should it not give 3:00 for the input 3:00 3:00 3:00. When x=0 :P – Lakshay Garg Jun 3 '15 at 19:53
• @LakshayGarg no. Problem is they each give a different time so x cannot be 0. – edc65 Jun 3 '15 at 20:11
• With each upvote you get, you should edc65++. – Alex A. Jun 4 '15 at 18:28

# Python 3, 166 163 bytes

L=input().split()
f=lambda x:int(x[:2])*60+int(x[3:])
s=""
for a in L:s=[s,s and"Look at the sun"or"The correct time is "+a][(sum(map(f,L))-3*f(a))%720<1]
print(s)


Uses

      a-b == b-c
<==>  a+c-2*b == 0
<==>  (a+b+c)-3*b == 0


with arithmetic being minutes modulo 720.

• Using a similar approach (a+a-b-c==0) I found that sorting is not needed. Did you try to avoid it? – edc65 Jun 3 '15 at 17:07
• @edc65 I did try to keep a count instead but I think it turned out longer. I also had a+a-b-c initially but that's a lot of calls to f (and pre-list-comp is long) – Sp3000 Jun 3 '15 at 17:13

# Python 2, 254 ... 199207203194 200 Bytes

Probably a few ways to shorten this, give me some time..

t=[x.split(':')for x in raw_input().split()]
a,b,c=[int(x[0])*60+int(x[1])for x in t]
print["The correct time is "+':'.join(t[[(c+a)%720%(b/2)<1,2][(a+b)%720%(c/2)<1]]),"Look at the sun"][(a-b)%240<1]


Thanks to Sp3000 helping me fix this.