Reasoning does the trick .
From your provided equation , try to get to the following result -
( b - c ) ( a - c ) = c 2
Let " P " be a prime which divides " c " .
Then obviously , " P 2 " will divide " c 2 " , i . e , " P 2 " will divide " ( a - c ) ( b - c ) " . But , if
" P 2 " divides both of them , then there will arise a contradiction --
since , a - cP = aP - cP <----- this must be an integer , so " P " must divide " a " , as our early
assumption was that " P " divides " c " . But " P " cannot divide both of them , as they don ' t
have any common factor . So " P 2 " will either divide " ( a - c ) " or " ( b - c ) " .
Hence , either " ( a - c ) " or " ( b - c ) " is a perfect square . But if one is a perfect square ,
then the other must also be a perfect square , since their product is a square = c 2 .
Now , choose , a - c = k 2 , and b - c = c 2k 2 , where obviously " k " divides " c " .
So , ( k + ck ) 2 = k 2 + 2 c + c 2k 2 = a - c + b - c + 2c = a + b
As , " k " divides " c " , hence " ( k + ck ) 2 " must also be an integer , rather a square of an integer .
So , " a + b " is an integer which is a perfect square .