My thoughts, ideas and processes with computing

Posted: 2019-01-15 | Last updated: 2019-01-15

As a bit of fun on a Saturday afternoon I've created a small program to convert a decimal number to hexadecimal (base 10 to base 16).

So base 10 is the decimal numbers we all use day to day.

Here is 0 to 15:

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15...]

Base 16 is hexadecimal, mostly commonly used (for me at least) in CSS for web development. If I wanted to change to colour of something to white I might use:

color: #fff;

or to make something black:

color: #000;

Here is 0 to 15 again but this time in hexadecimal:

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F]

So loosely speaking, this is Cascading style sheet's (CSS) way of saying colour this element's pixels the defined colour mixture in RGB (red, green, blue)*. Using the example from before;

"FFF" is saying use the colour mixture of 15, 15, 15 for these pixels. Since we are working in base 16; 0 to F is 16 values long (go ahead, count them above). 15 is the highest possible value in base hexadecimal for a single component and so what it is really saying is colour these pixels with 100% of each colour (red, green, blue). This equates to white. "000" similarly is 0% of each colour; black.

So for a very stark red I could use "F00", a green "0F0", a blue "00F".

**Commonly there are two components for each colour and sometimes one or two components for alpha (transparency) but I'll ignore those for simplicity.*

Before I'd decided which language to use I needed to understand the algorithm so I noted down the general method to take a number and convert it from decimal to hexadecimal. I know there are more efficient algorithms which can even convert from any base to another (I've seen them), but I wanted to write a short program to test my abilities.

Given a number in decimal.

- using integer division, divide the number by 16
- store the remainder as a component of the hexadecimal result
- if the result is 0:
- return the hexadecimal result

- else:
- go back to step 1 with the result as the number

A prime example of an algorithm which could benefit from recursion I thought.

I also noted that is the remainder was greater than 9, it would need converting to the corresponding letter. (Remember 10 in hex is 'A' or 'a').

I remembered that the ASCII number for 'A' is 65, 'B' is 66, 'C' is 67 etc. Therefore I devised a method for the conversion of components greater than 9 to the right letter in hex;

- if remainder is greater than 9:
- set the remainder to itself plus 55
- convert the remainder to the ASCII character at it's value

To demonstrate this; given the number 12, I want to end up with 'A' so: 12 + 55 = 67, the character at ASCII(67) is 'C'. Correct result!

I did a few calculations on paper to test my algorithm;

32 decimal as hex gave me #20 35 decimal as hex gave me #23 1526 decimal as hex gave me #5F6

So to cut to the chase, I decided to code this in python as a quick exercise and you can see the code below.

def to_hex(d): global hex if (d == 0): ##base case pass else: res = d // 16 rem = d % 16 if (rem > 9): rem = chr(rem + 55) hex.insert(0, rem) to_hex(res)

As I thought about the possibility of other algorithms to convert decimals to other bases I was intrigued that the base is used when converting to base 16:

res = d // 16 rem = d % 16

If I were to convert a number to base 2 (binary), would it simply be a case of changing the two numbers here to 2? Lets try.

Lets do an easy number to convert: 15

d = 15 #recursion 1 res = d // 2 ## 15 // 2 = 7 rem = d % 2 ## 15 % 2 = 1, so we have our first binary digit (1) #recursion 2 res = d // 2 ## 7 // 2 = 3 rem = d % 2 ## 7 % 2 = 1, Now we have our second binary digit (11) #recursion 3 res = d // 2 ## 3 // 2 = 1 rem = d % 2 ## 3 % 2 = 1, our third binary digit (111) #recursion 4 res = d // 2 ## 1 // 2 = 0 rem = d % 2 ## 1 % 2 = 1, our fourth binary digit (1111) #recursion 5 d is now 0 so we stop with the result: 1111

So now we have what looks like a binary number, lets just check it is 15 by converting 1111 back to decimal:

Decimal 8 4 2 1 Binary 1 1 1 1 8 + 4 + 2 + 1 = 15

Adding a few 0's to the nibble to make it a full byte: 0000 1111 = 15

So it would appear to work. More checks would be needed before modifying my function but it would be a relatively simple case of introducing a new variable parameter called base which would be used to make the conversion.