Data basics, clearly explained

Structure and meaning

Data format

A set of rules for representing information. JSON, CSV, hexadecimal text, and URL components use different rules and suit different purposes.

Syntax

The grammar of a format: which characters, separators, quoting rules, and structural arrangements are allowed. Valid syntax alone does not prove the data is correct.

Parse / parsing

Reading text according to a format’s grammar and building a structured representation from it. A JSON parser can detect malformed JSON, but it does not know your application’s requirements.

Serialization

Writing a structured value back into text or bytes in a chosen format. A serializer may change whitespace or presentation while representing the same parsed value.

Schema

A machine-readable description of expected data structure and constraints, such as required JSON properties, permitted types, and value ranges.

Validation

Checking data against a defined rule set. Syntax validation, schema validation, business-rule validation, and factual verification answer different questions.

Delimiter

A character that separates fields or values, such as a comma or tab in a delimited table. It may need quoting or escaping when it occurs inside a field.

Dialect

A particular set of conventions used by a format, often covering delimiter, quote character, escaping, line endings, encoding, and headers. CSV files commonly differ in dialect.

Header row

A table row containing column names instead of ordinary records. Whether a CSV begins with headers must be known before converting it into named JSON properties.

Text and bytes

Character

A unit of text as understood by a writing system. One displayed symbol can be represented by multiple code points or code units, so character count depends on how it is measured.

Unicode

A standard assigning code points to characters across writing systems. Unicode defines characters; a separate encoding represents them as bytes.

UTF-8

A widely used, variable-length encoding of Unicode code points as bytes. It is common on the web, but not every historical file or binary payload is UTF-8.

Byte

A group of eight bits, commonly used as a unit for stored data. Text must use an encoding to map characters to bytes.

Code unit

A fixed-size unit used by a particular text representation. JavaScript strings use UTF-16 code units; a visible symbol can occupy one or more of them.

Unicode normalization

A defined way of making certain equivalent text sequences use a consistent representation. It can change the underlying code-point sequence, so apply it only when the destination expects that normalization.

Line ending

The byte or character sequence marking a new line. LF and CRLF are common conventions; some tools normalize between them.

Conversion and checking

Encoding

Representing information in a defined form, such as Base64 or hexadecimal. Encoding is generally reversible and does not provide secrecy.

Encryption

A cryptographic process intended to protect confidentiality using an algorithm and key. It is not the same as Base64, hexadecimal, escaping, or obscuring text.

Escaping

Replacing characters with a representation that can be safely interpreted within a particular syntax or context. HTML escaping is not universal sanitization, and URL escaping is not authorization.

Normalization

Transforming values into a chosen consistent form, such as trimming spaces or standardizing line endings. Normalization can discard meaningful distinctions.

Lossless / lossy conversion

A lossless conversion preserves all distinctions relevant to the chosen definition of equivalence. A lossy conversion intentionally or incidentally discards information, such as type or precision.

Round trip

Converting data from one representation to another and back. A round trip may preserve meaning without reproducing the original bytes, whitespace, types, or order.

Precision

The detail with which a value can be represented. Some numeric types cannot exactly hold every large integer or decimal, so identifiers and exact quantities may need string or precision-aware representations.

Checksum / hash

A fixed-length value derived from data, often used to detect changes or identify content. A hash is not encryption and should not be treated as a proof of source authenticity unless paired with a trusted signature or authenticated process.

Privacy and workflow

Client-side processing

Code runs within the user’s browser rather than sending the transformation to an application server. This reduces one kind of data transfer but does not control extensions, device software, network observation, or host logs.

Data minimization

Using only the information necessary for a defined task and keeping it only as long as needed under the applicable policy.

Personal data

Information relating to an identified or identifiable person under applicable law. Definitions and obligations vary by jurisdiction and context.

Data provenance

Information about where data came from, how it was changed, and which process produced a result. Provenance supports reproducibility and review.

Audit trail

A record of relevant actions or changes retained in an approved system. FormatInstant does not create a transformation history; organizations needing one should use their designated audited workflow.

Want to apply these terms? Read the step-by-step guides, follow the working guidelines, or use the data readiness survey to think through a workflow before entering information.