Thursday 26 January 2017

Understanding Base64 Encoding #2

Tier 2

If Tier 1 is about establishing the scantest familiarity with a subject – hoping to avoid looking glassy-eyed whenever it’s mentioned – then Tier 2 is about beginning to understand the topic; perhaps a cursory interest has been kindled and/or you’d like to be able to do a bit more than just identify the subject by sight.

To that end, one of the first questions I like answered when getting to grips with a new topic is “why does this thing exist?”. I’m going to begin to attempt to answer this question for base64 encoding by giving a disingenuous, rather long-winded, somewhat tortured analogy. I promise I’ll make amends in later tiers.

Imagine a strange parallel universe in which inter-computer communication has never happened. The parallel universe’s computers work in the same manner as ours, just no one ever bothered to invent the technologies which allow computers to communicate: no Internet, Bluetooth, portable digital devices – no floppy discs, CDs, DVDs, USB drives, etc. Essentially, each computer is a lonely digital island.

In this reality, if I create a super-cool bitmap image in the alternative universe’s version of MS Paint, you’d physically have to come over to my house and look at it on my screen; I have no digital means by which to transmit the data to you. To add to my misery, you live on the other side of the country and, despite my enthusiasm and entreatment for you to come visit, you’re not going decamp for the sake of one bitmap image.

So, scratching my head, I begin to think about the problem and in a fit of pique I come up with my first – and worst – solution to this problem: I’m going to write the binary code out on pieces of paper and send the code in the post to you. Every single one and zero. And then when you receive the paper full of bits you can key them all in at your end and recreate the image. Perfect!

However, I soon find, even if I only wanted to send the small 10 x 10 pixel image from Tier 1 it’s ~1000 bytes. And given there are 8 bits in a byte that’s ~8000 ones and zeros I’ll have to transcribe! I’m not so keen on this and imagine you’re even less keen about having to key 8000 binary digits in at your end. We need a shortcut.

I’m convinced the part about mailing you the code still has merit but I’m also certain that raw ones and zeros aren’t the answer. What I need is some sort of shorthand way of representing the same raw binary data; I need to encode it.

This is the essence of the problem base64 encoding looks to solve: how can a text-based medium, in our case pieces of paper, be re-purposed to effectively transmit binary data.

Tier 3 will, hopefully, begin to straighten this all out…

Next >> Understanding Base64 Encoding #3

Tuesday 24 January 2017

Understanding Base64 Encoding #1

Disclaimer: I’m writing this blog post in an attempt to present a tiered approach to learning a new subject. It's also to solidify my understanding of the topic of base64 encoding as well as to act as an aide-memoire. I’m not presenting this information as infallible fact.
Preamble: Personally, learning a new programming concept (or any complex topic for that matter) requires me to take a very particular approach if I want gain and maintain a comprehensive understanding of it, and I don’t see resources which represent and facilitate my learning process very much in evidence.

Learning for me involves moving from the general to the specific and for my sources of information to assume as little as possible while establishing context and purpose quickly. Producing this type of learning resource usually manifests in tiered levels of explanation. To my mind, Tier 1 is where the biggest shortage of good resource on a topic generally is. It should be what the opening paragraph of the Wikipedia topic strives to attain: a succinct and clear overview of the topic that someone immersed in the relevant field can read and feel more illuminated right away. Further tiers of explanation should elaborate on what previous tiers have established.

Let me try presenting the first couple of tiers for base64 encoding in the style I'm talking about.

What I assume: you have a programming background and that you’re looking to better understand base64 encoding.

Tier 1:

Okay, Tier 1 explanations might be relevant if you’ve just heard someone say “base64 encode” in a meeting and you’re thinking “I should probably have some idea what on Earth they’re talking about”; you’re googling about for five minutes to see if you can shed some light on the topic.

Wikipedia’s Base64 opening salvo is: “Base64 is a [...] binary-to-text encoding scheme that represent[s] binary data in an ASCII string format”.

This isn’t particularly illuminating on its own but there are a couple of clues in there: it’s something do with binary data being represented as ASCII characters.

Warning: rather unhelpfully, it is possible to immediately jump down the rabbit hole with base64 encoding and you may be thinking, as I was, “hang on a minute, everything eventually boils down to binary data - including ASCII characters - so that seems like a bit of a nonsense”. Or perhaps you have come across an example whereby someone is showing you how they converted a sentence (one string of characters) into base64 encoded text (another string of characters) and are thinking “what could possibly be the value in that!?”. If you’ve done either (or both) of these things, please, for the moment, put those thoughts on ice – don’t worry, I’m with you comrade, I feel your pain.

A concrete example might help. Imagine I have an 10 x 10 pixel jpeg image (some binary data) and I want to represent it (for some ungodly reason) as ASCII characters. Up steps base64 encoding. In fact, here is a base64 encoded 10 x 10 jpeg:


Sceptical? If you copy that text and save it into a new text file (called, say, “encodedJpg.txt”) and then navigate to the folder the file is saved in from a Windows command prompt, you can run the following command certutil -decode encodedJpg.txt 10x10.jpg and you should see the jpg recreated.

You can turn the jpg back in to the text above by running the alternative certutil -encode "input" "output" command.

And that’s Tier 1. For the moment we’re not going to worry about the mechanics of the operation, it’s enough to know that base64 encoding changes binary data into text that looks like the above. Why you'd want to do such a thing and how it's achieved are Tier 2 explanations. N.B. the binary data doesn’t have to be a jpeg image, it could be anything: an executable, a zip file, a Word document, etc.