How is SQL pronounced, and which form is most common among database professionals?
SQL — two accepted pronunciations.
"sequel" /ˈsiːkwəl/ — the historical, very common spoken form (SQL descended from a language called SEQUEL)
"ESS-QUE-ELL" /ɛs kjuː ɛl/ — the spelled-out form, equally correct and arguably more "official"
Both are universally understood. Notice that product names follow conventions: MySQL is officially "my-ess-que-ell" but many say "my-sequel"; PostgreSQL is "post-gress-que-ell" (often shortened to "Postgres"); NoSQL is "no-sequel" or "no-ess-que-ell". Collocations: write a SQL query, a SQL injection, raw SQL. Never say "squeal".
2 / 5
How is UUID (and the related GUID) pronounced by developers?
UUID is spelled out; GUID is usually a word.
UUID — "YOU-YOU-EYE-DEE" /juː juː aɪ diː/ (Universally Unique Identifier). The double "U" is read as two separate "you" sounds.
GUID — most commonly "goo-id" /ˈɡuːɪd/ (two syllables), Microsoft's term for the same concept (Globally Unique Identifier). Some say "gwid" /ɡwɪd/ as one syllable.
Collocations: generate a UUID, a v4 UUID, the record's GUID. Avoid "guide" — that is a different English word. UUID is the cross-platform/standards term; GUID is the same idea in the Microsoft/.NET ecosystem.
3 / 5
How are JSON and CSV pronounced?
JSON becomes a word; CSV stays an initialism.
JSON — "JAY-son" /ˈdʒeɪsən/, exactly like the name Jason. (Its creator Douglas Crockford notes either "JAY-son" or "jay-SAHN" is fine, but "JAY-son" dominates.) JavaScript Object Notation.
CSV — "SEE-ESS-VEE" /siː ɛs viː/ (Comma-Separated Values). The letters "csv" have no vowel, so it cannot form a word.
Collocations: parse the JSON, a JSON payload, valid JSON, export to CSV, a CSV file. Related: YAML is "YAM-ul", XML is spelled out "EKS-EM-ELL", TSV is "TEE-ESS-VEE".
4 / 5
How are the analytics/transaction acronyms OLAP and OLTP pronounced?
OLAP is a word; OLTP is spelled out.
OLAP — "OH-lap" /ˈoʊlæp/, said as a word (Online Analytical Processing). The "lap" syllable works naturally.
OLTP — "OH-ELL-TEE-PEE" /oʊ ɛl tiː piː/, spelled out (Online Transaction Processing) because "oltp" cannot form a syllable.
This is a classic asymmetric pair: the same prefix, but one ends in a pronounceable cluster and one does not. Collocations: an OLAP cube, OLAP workloads, tune the OLTP database, OLTP vs OLAP. Related data terms: ETL "EE-TEE-ELL" (Extract, Transform, Load) and BI "BEE-EYE" (Business Intelligence) are both spelled out.
5 / 5
How are the data-design acronyms ACID, CRUD and ORM pronounced?
ACID and CRUD are words; ORM is spelled out.
ACID — "acid" /ˈæsɪd/, exactly like the chemistry word (Atomicity, Consistency, Isolation, Durability). Collocation: ACID transactions, ACID guarantees.
CRUD — "crud" /krʌd/, like the informal English word (Create, Read, Update, Delete). Collocation: a CRUD app, basic CRUD operations.
ORM — "OH-ARE-EM" /oʊ ɑːr ɛm/, spelled out (Object-Relational Mapping/Mapper) because "orm" is awkward as a syllable. Collocation: use an ORM, the ORM layer.
The pattern again: pronounceable letter strings become words, awkward ones stay initialisms.