create a URL shortener with Base 62?

Question

I understood the process to shorten the URL with base 62 at How do I create a URL shortener?.

Steps given are

1. Think of an alphabet we want to use. In your case, that's [a-zA-Z0-9]. It contains 62 letters.
2. Take an auto-generated, unique numerical key (the auto-incremented id of a MySQL table for example).
3. For this example, I will use 12510 (125 with a base of 10).
4. Now you have to convert 12510 to X62 (base 62)

My question is why not just create unique numerical key and return it ? What is the advantage of concerting numerical key > Base 62 > then Finally some alphanumeric number ? Is it because final alphanumeric number will be much smaller than unique numerical key ?

Answer 1

I can see you understand that, yes, a number written in base 62 takes less characters than a number in base 10 just like a number in base 10 takes less characters than a number in base 2 (e.g. 0101 is 3 characters longer than just '5').

So, I'll answer specifically "Why".

Sometimes a link is shortened to be more visually pleasing. A company worried about their public perception likely doesn't want their links to look like an error code due to how long they are so they resort to shortening. That's why some url shortening services allow you to add your own "vanity url" which customizes the domain name, so that a link can be shortened and branded.

Other times a link is shortened to minimize character count when working with constraints, like Twitter. For example, at my company we shortened the links in our automated Twilio messages because SMS messages that contain more than 160 characters are technically 2 concatenated messages so it is more expensive to send.

And finally if the link is being shared through a medium that cannot be directly clicked on (e.g. verbally, on paper), making it shorter makes it much easier to type into an address bar manually. (Imagine trying to type the url to this SO question when someone is reading it to you.) I assume this is also at least partially why the base used for these links usually stop at around 62. If you start including other arbitrary characters to higher the base and consequentially make the link marginally shorter, it'll become harder to communicate, read and type. ("domain.name/5omeC0d3" vs "domian.name/🈲}♠ "

Answer 2

Yes. The idea is to make it short and usable in a URL. A number in base 62 will use fewer characters than the same number in base 10. Notice also that URL shorteners use short hosts, such as g.co.