01/ Definition Term

DEFINE: Tabular Data in CLNv2.3 NEW

=Tabular data — two-dimensional, fixed-column, scalar-cell information =new in CLN v2.3 — the missing structure between lists and prose .one header row: ((( col | col | col ))) .data rows: (( val | val | val )) .cells separated by | pipe character .empty cell = whitespace between pipes →renders as HTML table — reads as plain text✓ portable

COMPARE: Before and after v2.3motivation

// Before (v2.2 — no tables) -metric comparisons buried in nested lists -column alignment lost in raw text -reader can't scan vertically across options // After (v2.3 — tables) +column schema declared once in header row +vertical scan path: compare column values at a glance +still plain text — works in Slack, email, terminal +|| spoilers || and »links permitted inside cells

02/ Syntax Grammar

SYNTAX: Table grammar — plain text sourcecopy-paste ready

// Minimal example

((( Metric | Option A | Option B ))) (( Speed | fast | slow )) (( Memory | O(1) | O(n) )) (( Stable | No | Yes ))

// Full document context

ANALYSIS: Algorithm comparison = choosing sort algorithm for 10M records ((( Algorithm | Time | Space | Stable ))) (( QuickSort | O(n log n) | O(log n) | No )) (( MergeSort | O(n log n) | O(n) | Yes )) (( HeapSort | O(n log n) | O(1) | No )) → MergeSort chosen — stability required

// Empty cell = whitespace between pipes

((( Feature | Free | Pro | Enterprise ))) (( Storage | 5GB | 50GB | Unlimited )) (( SSO | | | ✓ )) (( API access | | ✓ | ✓ ))

DEFINE: Token anatomyparser view

=header token: ((( … ))) =data token: (( … )) =cell delimiter: | pipe, one or more spaces either side .parser reads: ((( → header → ))) .then: (( → row → )) until no more rows .column count inferred from header !header row must precede all data rows !do not nest tables inside tables !do not use ( ) blocks inside cells

03/ Rules & Constraints Spec §7.2

SPEC: Complete rule set — CLN §7.2normative

// Structure .exactly one ((( ))) header row per table .one or more (( )) data rows follow immediately .all rows indented at same level as the header .table ends at first non-table line at same indent // Cells =cell content: scalar — one value per cell =leading/trailing whitespace trimmed from each cell =empty cell = one or more spaces between | pipes

// Permitted inside cells +|| spoilers || — deferred cell content +»[text]##id cross-references +Unicode characters, emoji, symbols // Forbidden inside cells !nested tables ((( ))) !parallel blocks ( ) !multi-line cell content !CLN operators as cell content (= . etc.)

04/ Table vs Parallel Blocks Decision

? Fixed column schema across all items?

Yes → continue →

No → use ( ) blocks

→

? Cells are scalar — one value each?

Yes → continue →

No → use ( ) blocks

→

? Reader scans vertically (compare columns)?

Yes → continue →

No → use ( ) blocks

→

→ Use Table

((( col | col )))
(( val | val ))

EXAMPLE: Table — API tiers comparisonscalar cells ✓

=comparing three API rate-limit tiers for pricing page

Tier	Req/min	Burst	Price/mo	SLA
Free	60	10	$0	—
Starter	600	100	$29	99%
Pro	3 000	500	$99	99.9%
Enterprise	Custom	Custom	Custom	99.99%

*fixed schema · scalar cells · vertical comparison — table is correct choice

EXAMPLE: Blocks — architectural optionshierarchical ✓

=cells need multi-line + nested operators → use ( ) blocks, not table

= Monolith

+simple ops · fast start -scale ceiling !deploy = full redeploy

. Microservices

+independent scale -ops complexity ↑↑ ~flexibility vs coherence

*each block has operators, subpoints, tensions — not scalar → blocks win

05/ Live Examples Real-World

COMPARE: Database enginestech selection

=choosing DB for analytics workload · 500M rows

Engine	Query	Write	Joins	Cost
Postgres	⚠ slow	✓ fast	✓ full	$
ClickHouse	✓ fast	⚠ batch	✗ limited	$$
BigQuery	✓ fast	⚠ stream	✓ full	$$$
DuckDB	✓ fast	⚠ local	✓ full	free

→DuckDB for local analysis · BigQuery for prod✓

ANALYSIS: Sprint velocitymetrics

=Q1 2026 · team of 4 · 2-week sprints

Sprint	Points	Done	Bugs	Delta
S1	34	28	3	-6
S2	36	35	1	-1
S3	38	38	0	±0
S4	40	42	0	+2

→velocity trend: +2.9 pts/sprint average !S1 deficit caused by onboarding new engineer

COMPARE: PKM tool matrixfeature table

=evaluating knowledge tools for team of 12

Tool	Local	API	CLN	Price
Obsidian	✓	plugin	✓	free
Notion	✗	✓	⚠	$16/mo
Logseq	✓	beta	✓	free
Plain CLN	✓	✓	✓	free

*CLN files work in any tool — format is portable

06/ Inline Features in Cells Spoilers · Links

GUIDE: Table cells with spoilers and cross-references|| || and »links inside (( ))

// ##tbl-inline Spoilers inside cells — click to reveal

Algorithm	Best case	Worst case	Stable?	Verdict
QuickSort	O(n log n)	O(n²)	No	avoid for sorted input — use randomised pivot
MergeSort	O(n log n)	O(n log n)	Yes	preferred when stability matters — extra O(n) space cost
TimSort	O(n)	O(n log n)	Yes	Python/Java default — optimised for real-world data patterns
HeapSort	O(n log n)	O(n log n)	No	in-place but poor cache behaviour — niche use

// »links inside cells

Task	Depends on	Status
Tokenizer core	—	✔ done
Addressing system	»[Tokenizer]	☐ wip
Table parser	»[Addressing]	☐ next
JSON export	»[all above]##tbl-inline	☐ planned

07/ Self-Test || spoilers ||

QUIZ: Tabular data in CLN v2.3click to reveal answers

*Q: What token marks a table header row? ((( col | col | col ))) — triple parentheses *Q: What token marks a data row? (( val | val | val )) — double parentheses *Q: How do you create an empty cell? whitespace between pipes: (( val | | val )) *Q: Can you nest a table inside a table? NO — strictly forbidden. Flat structure only. *Q: Which inline tokens ARE permitted inside cells? || spoilers || and »[text]##id cross-references *Q: When should you use ( ) blocks instead of a table? When cells are hierarchical, multi-operator, or non-scalar — i.e. anything you wouldn't put in Excel →three questions: fixed schema? — scalar cells? — vertical scan? →all yes → table. any no → parallel blocks.✓ rule

---