Search the site’s resources

Topic or keyword

Content familyAll content familiesDossierProtocolGuidePublic sourceScenarioToolPlatformServices Search

All words must match, regardless of case or accents. Search stays in the selected language.

From two characters: use arrows to choose, Enter to open, Escape to close. You can also submit the full search.

1585. Guide

      Axis order: identify both CRS and reading contract

      Plausible numbers do not locate a point without CRS, units and software conventions. GDAL and OGC API Features explain why WGS 84 alone does not settle data order.

1586. Guide

      GeoJSON extent: five degrees or almost the world?

      RFC 7946 supplies two extents using the same longitude extremes: 5° across the antimeridian versus 355° the other way. Sorting west and east changes the selected region. The explorer uses these documented examples.

1587. Guide

      Geodesic distance: units, ellipsoid and argument order

      Official GeographicLib examples provide reproducible inputs and outputs on WGS 84. These documented calculations do not route or geocode your addresses.

1588. Guide

      Spatial joins: boundary points can change the result

      Predicate choice changes matches for identical coordinates. The first three rules concern 2D geometry. The illustrative polygon, hole and points represent no actual commune.

1589. Guide

      Unicode normalization: decide what a join may equate

      Visually identical labels can encode different sequences. UAX #15 defines four forms. Ten official cases below come from Unicode 17.0.0 conformance data, explicitly versioned for reproducibility; this does not identify a browser’s Unicode version.

1590. Guide

      Text length: bytes, code points and visible signs differ

      Column length needs a unit. Counts below follow UTF-8 and UTF-16 for four exact sequences, not a user file. One perceived sign may use several code points. Table notation: cp = code points; B = bytes.

1591. Guide

      CSV: a displayed line is not always a record

      Informational RFC 4180 describes a common comma format, not every dialect. Splitting on every comma and newline can destroy cells. Fragments here illustrate syntax, not actual observations.

1592. Guide

      Empty cells: CSV syntax alone does not determine meaning

      W3C CSVW separates string value and semantic value. Quoting an empty string is not a data distinction in that model. These are CSVW rules, not universal SQL or spreadsheet import behavior.

1593. Guide

      JSON: valid text can still lose an identifier

      RFC 8259 defines exchange syntax, not unlimited precision in every reader. Long numeric identifiers may change during parsing, before a spreadsheet display. Examples are teaching forms, not assigned identifiers.

1594. Guide

      Spreadsheet dates: serials need the workbook’s date system

      Microsoft documents 1900 and 1904 systems. Its 5 July 2011 example uses two serials for one date. Inspect workbook settings before comparing extracts; display formatting cannot supply that setting.

1595. Guide

      Outer joins: ON and WHERE do not retain the same rows

      PostgreSQL 18 publishes two three-row teaching tables. The explorer reproduces them for nine join/filter combinations, without real entities or database access. Counts apply only to this published example.

1596. Guide

      SQL NULL: unknown is not another spelling of false

      SQL has three-valued logic. These comparisons and truth-table rules are documented in PostgreSQL 18, independent of any dataset. They explain discarded outer-join rows.

1597. Guide

      Keys and constraints guarantee different things

      An id column is not automatically a primary key. Effective, validated constraints define guarantees. These rules are for PostgreSQL 18; other systems may treat unique-constraint NULLs differently.

1598. Guide

      After a join: count rows, present values or distinct objects

      A total can change because its counting unit differs. PostgreSQL distinguishes input rows, non-null values and distinct values. The explorer shows all three without treating teaching IDs as statistical measurements.

1599. Guide

      JSON Schema: present, non-null and allowed are separate checks

      This reference targets dialect 2020-12, which the validator must support. Fragments are pedagogical. Object rules also assume type: object; required alone does not make arbitrary values objects.

1600. Guide

      PROV provenance: connect version, process and responsibility

      W3C PROV-O describes how documents or data were produced. Arrows follow property direction, not a left-to-right timeline. E0, A and E1 are teaching labels, not real resources.

1601. Guide

      Confusion matrices: identify axes before naming errors

      Automatic citation or answer checks compare a decision with a reference annotation. Separate detection mistakes, annotation mistakes and swapped columns. scikit-learn puts reference labels in rows and predictions in columns; other displays may reverse them. Positive is fixed to 1 here.

1602. Guide

      Precision, recall and false-positive rate use different denominators

      These unweighted binary formulas use a declared positive class. They can evaluate automatic presence annotations, without treating a brand mention as an accurate citation or useful outcome.

1603. Guide

      Change the threshold, keep the same four scores

      The published average_precision_score example has references [0,0,1,1] and scores [0.10,0.40,0.35,0.80]. This explorer uses those exact teaching pairs. Positive means score at least the selected threshold; these are not real agent results or guaranteed calibrated probabilities.

1604. Guide

      ROC AUC and average precision are different ranking summaries

      The official four-pair AP example yields ROC AUC 3/4 and AP 5/6. These are calculations on teaching data, not measured agent performance. They do not change when only the decision threshold changes.

1605. Guide

      Macro, micro and support weighting: five report cases

      Published classification_report compares [0,1,2,2,2] with [0,0,2,2,1]. Scores reproduce two-decimal output; original pairs remain below. This three-class teaching example is separate from the four-score binary explorer.

1606. Guide

      Independent evaluation: splitting cases is not enough after answers influenced tuning

      A score can be optimistic with correct formulas. scikit-learn distinguishes training, tuning and final evaluation. An agent has the same issue when instructions change after repeatedly inspecting a supposed generalization test.

1607. Guide

      Read Core Web Vitals thresholds without inventing a verdict

      LCP, INP and CLS describe three different dimensions of experience. Published thresholds classify a value; they do not turn an isolated test into a representative observation of every visitor.

1608. Guide

      CrUX: separate the window, scope and available observations

      A CrUX response describes an eligible population within a rolling window. Before comparing two p75 values, retain the requested key, devices and returned dates with the result.

Choose a starting point

Choose a method

Start from a task: verify, compare, document or decide.

Check a source

Examine the origin, date, scope and limits of information.

Understand a term

Find definitions before applying a protocol.

How can I refine the search?

Start with the topic

Enter one or two specific terms: citation, CSV or provenance. Every word must be present, so a long sentence can exclude useful pages.

Broaden before filtering

Search all families first. Then select a guide, protocol or source according to the kind of answer you need.

Try different wording

Try a synonym or a term in the selected language. Quotes, OR and asterisks are not search operators.