Search

Search the site’s resources

A single starting point to find the method, source or situation relevant to your question.

Find a method for your task

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.

1682 results

Clear filters
  1. Guide

    Statistical unit / Data grain

    The statistical unit is the object observed; grain specifies what a row represents and its dimensions. Several rows can describe one unit.

  2. Guide

    Missing value / Imputation

    A missing value lacks a usable observation; imputation assigns an estimate using a method. That estimate must remain identifiable.

  3. Guide

    Statistical stock / Statistical flow

    A stock describes a position at a date; a flow describes events over a period. Adding successive stocks does not produce the period’s flow.

  4. Guide

    Percentage point / Relative change

    A percentage-point change subtracts two rates; relative change divides the difference by the initial rate. State the base and the result’s unit.

  5. Guide

    Calendar date / UTC

    A civil date names a calendar day; a UTC timestamp identifies an instant. Converting a date without a time or zone would invent absent precision.

  6. Guide

    HTTP 202 — request accepted / HTTP 206 — partial content

    202 means processing was accepted but not completed; 206 means a response supplies requested partial content. Neither replaces checking the expected outcome.

  7. Guide

    Private cache / Shared cache

    A private cache serves one user; a shared cache can serve several users. Purging or bypassing them involves different operations.

  8. Tool

    Start from zero without inventing a percentage

    Going from 0 to 12 adds 12 units. Relative change and the base-100 index remain undefined: division by a zero baseline cannot yield an interpretable percentage.

  9. Tool

    Read a decrease with its sign

    Going from 120 to 90 gives −30 units, or −25%. The final index is 75 with an initial base of 100; that is neither a 75% fall nor a loss of 25 units.

  10. Tool

    Changing the base changes the percentage

    From 80 to 100, the increase is 25%. Returning from 100 to 80 is a 20% fall because the denominator becomes 100. The difference remains 20 units in absolute magnitude.

  11. Tool

    A small point increase can be large proportionally

    From 0.2% to 0.3%, the gap is 0.1 percentage point and the relative increase is 50%. This arithmetic proves neither causality nor significance; counts and the observation design are still needed.

  12. Tool

    The larger group changes the overall rate

    Two disjoint, compatible groups have 9 cases out of 10 and 10 out of 1,000. Pooling gives 19 out of 1,010, approximately 1.88%. The simple average of 90% and 1%, which is 45.5%, does not describe this population.

  13. Tool

    Overlapping groups stay separate

    The same 100 people are observed on two dates, with 10 and then 20 cases. Separate rates can be calculated, but pooling them as 200 distinct people would count individuals twice. Compatibility is therefore not confirmed here.

  14. Tool

    Filled fictional excerpt · Crawling and indexing audit

    Retain successful access and exclusion from indexing as separate observations. Status 200 does not close this review. After changing the directive, inspect the final page and separately check any available indexing status.

  15. Tool

    Filled fictional excerpt · Evidence and source register

    Limit the conclusion to the 100 reviewed records; the other 200 remain unknown. Copies are not additional collection. Obtain the selection rule and a distinct observation covering unreviewed records.

  16. Tool

    Filled fictional excerpt · AI citation observatory

    Separate citation presence from support for the claim. Do not count this as a verified answer. Read the final destination, retain the passage and repeat the question under documented conditions.

  17. Tool

    Filled fictional excerpt · Control AI crawler access

    The fictional crawler receives a disallow rule, but the resource responds without authentication. Do not treat the rules file as access protection. Test server-side protection with an unauthenticated client, then inspect the relevant crawler rules.

  18. Tool

    Filled fictional excerpt · Build a verifiable entity profile

    Retain this entry as an identified candidate in the case. Another result with the same name must not be merged without matching its attributes. Compare identifier, jurisdiction, relevant address and period before combining records.

  19. Tool

    Filled fictional excerpt · Audit a multilingual website

    The expected Spanish variant serves English and points to another language. Keep both defects separate rather than marking the translation checked. Check text, declared language and reciprocal links after correcting the Spanish version.

  20. Tool

    Filled fictional excerpt · Map AI system dependencies

    The fallback is not operational yet. Retain the critical dependency and restrict permissions to the authorised task. Run a limited test using replacement access; record elapsed time and unavailable functions.

  21. Tool

    Filled fictional excerpt · Audit structured data

    Markup and the visible date of the same event disagree. Valid syntax does not validate the stated fact. Check the date against its source, then inspect visible content and final markup together.

  22. Tool

    Filled fictional excerpt · Evaluate an AI agent before release

    No change does not mean the test passed: the attempted action exceeds the read-only mandate. Repeat after correcting boundaries; retain the proposal, rejected call and final state.

  23. Tool

    Filled fictional excerpt · Maintain a reliable editorial corpus

    Mark the passage for review without guessing a replacement date. Retain previous wording and the review reason. Check the relevant notice, effective date and language versions before approving the correction.

  24. Tool

    Filled fictional excerpt · Audit accessibility across a web journey

    The tested journey remains blocked for keyboard use even if other fields work. Locate the control and reproduction path. Repeat the whole journey with a keyboard after correction, including validation and error feedback.

Choose a starting point

Check a source

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

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.

DEEPER READING / PRACTICE

Turn internal search into a useful user journey

A query produces several close results. Reading should help to distinguish the need definition folder, detailed method and implementation tool.

Updated on .

Priority checks

Formulate a specific need and then try neighboring terms. Read the descriptions and select the type of resource according to the searched step: understand, run or control.

Interpretation pitfalls

The first result is not always the best document to the context. An absence can come from the vocabulary used rather than a lack of content.

Working record to retain

A short route connects a need definition page, a guide and a checklist. Maintaining unresolved questions helps to rebuild research.

PRACTICAL CASE / WORKED EXAMPLE

A natural requirement does not match the exact title

The following case illustrates an approach to defining the need. Situations are pedagogical: they serve to examine the choices and conditions of success, without describing a fulfilled mission or announcing a guaranteed result.

Example context

A reader is looking for how to check a number, while the resources use the word indicator. The results must allow to continue despite a different formulation of the title.

Suggested approach

Try the neighboring terms and read the descriptions. Differentiate methods of calculation, comparison and verification of sources. Choose the resource that corresponds to the actually sought-by-step.

Conclusion and next step

Research becomes a path of solution. The questions still open serve to clarify the next query instead of multiplying the words without clear intention.

Questions for deeper practice

How do you improve a search without losing the initial need?

Keep the starting question and then specify the object, operation and resource type. Useful results give a complementary vocabulary, but their title should not automatically replace the reader's target.

DETAILED WORKSHOP / APPLICATION

Refine a query for the task to carry out

  1. 01 / Prepare the necessary information

    Describe the initial question and the desired result. Identify the object, operation and type of resource that could help: file, method or tool.

  2. 02 / Carry out the approach

    Try a formulation and read the descriptions. Release the useful terms and test the variants without giving up the intention. Compare the resources found according to their contribution to the route.

  3. 03 / Examine an alternative situation

    If the results are close but too general, specify the operation or scope. A well-classed page is not necessarily the one that meets the searched step.

  4. 04 / Decide on the next step

    Research is useful when it leads to a resource whose response can be identified. Open questions are used to prepare for the next consultation.

  5. 05 / Hand over the verification records

    Keep a short track with request, resource and supplement. The reader can restart his work without repaired the tests that have led to useful content.

ANNOTATED WORKING DOCUMENT / REVIEW GRID

A track record from a request

The note retains question, tried formulations, useful results and next resource. It distinguishes the exploration vocabulary from the initial goal of the reader.

On a small screen, scroll the grid horizontally to read all three columns.

Example situations to assess
ObservationSupported interpretationNext step
A resource responds to the desired operation.The result can serve as a useful step in the course.Keep the passage and choose the necessary supplement.
The results remain general despite the good topic.The request must specify action, object or scope.Try an appropriate formulation without losing the initial question.
The first result has the word but not the answer.The presence of the term does not prove its relevance.Read the descriptions and compare the contribution of resources.

Reviewing the working document

The final verification verifies that each step reduces an open question. A link list is not yet a useful path if the reader cannot explain what each page brings.

REPORTING / EXAMPLE WORDING

Presentation of results: Turn internal search into a useful user journey

Point to clarify

A search result must be judged according to its contribution to the searched operation.

Before handing over

Check that the track keeps the initial question despite the vocabulary variants.

Example of a qualified conclusion

The resource selected meets the described stage of the need. The questions still open guidelines the next consultation and remain distinct from the answer found.