Writing about developer
Twelve write-ups on the formats that get pasted between tools all day: what base64 is for and what it costs, why YAML indentation bites, what is actually inside a JWT, when a v7 UUID beats a v4, and why a timestamp lands an hour off.
Why Your JSON Fails to Parse: the Usual Five
The parser tells you where it gave up, not where you went wrong. Five mistakes, the message each one produces, and a way to cut a stubborn file in half.
ReadWhat the = at the End of Base64 Means
The equals signs are bookkeeping for a last group that came up short. When they can go, and what a string’s length tells you before you decode it.
ReadThe Regex Mistakes That Waste an Afternoon
Three regex mistakes return something that looks almost right, which is why they cost hours instead of minutes. How to recognise each one in a minute.
ReadA Timezone Is Not an Offset
An offset is one number for one instant. A timezone is a set of rules that produces that number. Why the difference decides what you should store.
ReadWhy Putting a UUID in a URL Is a Tradeoff
A random identifier hides how many records you have and stops people counting their way through them. It also makes every URL unreadable and unsayable, and it is not access control.
ReadWhat Base32 Is and Where You Have Seen It
Five bits per character, a shrunken alphabet, and the reason your two-factor secret looks nothing like a JWT.
ReadWhat a UTM Parameter Actually Does
A UTM tag rides in the query string and nothing reads it on the way. Only the analytics script on the page it lands on ever looks at it.
ReadWhy JSON Has No Comments
Douglas Crockford took them out on purpose, and the workaround split into two dialects with the same idea and different names.
ReadYAML vs JSON: Which to Use, and Where Each Bites
JSON for what machines exchange, YAML for what people edit. The interesting part is the indentation, the type coercion and the country code that turns into false.
ReadShould You Use Base64 Images? When Inlining Helps
Inlining saves one request and costs you 33% more bytes, separate caching and lazy loading. The line sits at about a kilobyte.
ReadWhat Is Inside a JWT (and What It Does Not Prove)
Three base64url chunks, a handful of claims, and one signature that proves far less than most people assume.
ReadWhat Is Base64 and When You Actually Need It
Base64 turns bytes into 64 safe text characters, three at a time, and charges you a third of the file size for the privilege.
ReadWhat a Unix Timestamp Is and How to Read One
One number, counting seconds since 1970. How to read it at a glance, and why yours keeps landing in January 1970.
ReadUUID v4 vs v7: Which One to Use and When
v7 for database keys, v4 when the creation time is nobody else’s business. The bits behind that rule, and the two places v7 does nothing at all.
ReadWhat URL Encoding Is, and Why Your Link Breaks
Percent-encoding in plain terms: which characters need escaping, which never do, and the three ways it goes wrong.
ReadHTML Entities: What They Are and When You Need Them
An entity is a character written in plain ASCII. You need five of them, and the rest you can type straight into the file.
ReadHow to Format Minified JSON So You Can Read It
Re-indenting one enormous line is the easy part. The round trip through the parser is where things change without telling you.
ReadHow to Convert JSON to CSV and What It Loses
Every key becomes a column and every object becomes a row. The decisions start the moment something in your JSON is not flat.
ReadHow to Convert CSV to JSON (and What Breaks)
Rows become objects and headers become keys. The interesting part is the leading zeros, the stray quote and the nesting you cannot get back.
ReadBinary vs Hex vs Decimal: The Difference Explained
The base only decides how a quantity is spelled. Which one to use depends on whether a human or a circuit is reading it.
ReadWhere to start
If you only read two, make them what base64 is and what a Unix timestamp is. Both are things most developers use for years with a working but incomplete picture, and both produce bugs that are hard to spot precisely because the code looks right.
What is inside a JWT is the third, and it is the one with security consequences if the picture is wrong.
What these pieces have in common
These are mostly about formats, and a format is a set of decisions someone made long ago that you now have to live with. Why JSON has no comments is a design choice with a stated reason. UUID v4 against v7 is about a tradeoff between randomness and index locality that did not matter until databases got large. YAML against JSON is about readability bought with ambiguity.
The rest are the encodings that quietly sit between systems. URL encoding, HTML entities, base32 and number bases all solve the same shape of problem: getting data through a channel that reserves some characters for its own purposes.
Frequently asked questions
Why does JSON not allow comments?
Its designer removed them deliberately. Comments had been used to carry parsing directives in other formats, which turned a data format into something each parser interpreted differently. Leaving them out keeps JSON unambiguous. The common workaround is an ordinary key that readers ignore, such as a field named for a note.
Should I use UUID v4 or v7?
Use v7 when the UUID is a database primary key. It begins with a timestamp, so new rows land next to each other in the index rather than scattered through it, which keeps inserts fast as the table grows. Use v4 when you want no information revealed by the identifier at all.
What is the difference between base64 and base32?
Both represent binary data as text, with different alphabets. Base64 uses 64 characters including ones easily confused when read aloud or typed, and is more compact. Base32 uses 32 unambiguous characters, is about twenty percent larger, and is what two-factor setup keys use because people have to type them.
All categories
PDF 10Images 10Text 14Design 11Security 8Calculators 13Everyday 4Time 7