Editorial standards

What an article must add

An article appears here only if it gives a reader something they cannot get by skimming the current top results for its subject. That contribution has to be identifiable in a single sentence before the article is written. It might be a worked calculation, a reconciliation of sources that disagree, a diagnostic path where each branch is justified by a cited mechanism, an original mechanism diagram, a matrix of published process parameters, an interpretation of data someone else published, or original testing documented to the standard below.

Length is an output rather than a target. There is no word count, no citation count and no required section list, and there is no template: an article whose outline could be swapped with another article’s outline without either becoming incoherent is not ready to publish.

Articles carry no individual byline. They are written and edited by AnalogFixer, and the publication is answerable for every figure in them. We would rather be judged on whether the sources check out than on a name and a job title, which is a claim to authority that a reader cannot verify — and which is, in this subject, usually unnecessary: almost everything here traces back to a manufacturer document you can open yourself.

What each source class can establish

• Mechanism evidence. Peer-reviewed papers and published standards. Textbooks and monographs are secondary literature: good for orientation and for finding the primary work, but a specific mechanism claim traces to the primary source where one exists.
• Product specifications. Manufacturer technical data sheets and service manuals, cited with the exact product, formulation, revision and region, because these documents are replaced in place and change without announcement. They are authoritative for what a product is, and they are not an independent party.
• Safety and regulatory guidance. Safety Data Sheets, exposure limits and disposal rules. Required for any statement about handling, dilution, storage or disposal.
• Historical evidence. Contemporaneous documentation, patents and dated company material. A patent establishes what was disclosed on a date, not that the thing works.
• Practitioner observation. Community testing and trade writing, often the most current material that exists, attributed to the individual report it is, and never carrying a mechanism claim on its own.

How sources are shown in References

Every material claim is checked against the primary document before an article is published. The article’s References section lists each source with a working link, so a reader can check the article without asking anyone. Where a source has a revision or edition date, that date appears in the reference, because a number without that context cannot be used.

An article whose argument still depends on a claim that has not been verified against its source is not published. Not published here means absent rather than flagged: it has no page, it appears in no section listing, and there is no link to it. That is deliberate. Nothing unverified should be discoverable as though it were settled, and a warning notice is a weaker guarantee of that than simply not putting the page on the site.

Safety and documented-testing requirements

This is the commitment with consequences attached, so it is stated plainly. No statement about handling, dilution, storage or disposal appears on this site without a reference to the current product-specific Safety Data Sheet, identified by its revision date. Not the manufacturer in general, not a sheet for a similar product from another supplier: the document that applies to the material being described, as it currently stands.

Where a procedure carries a real hazard, the procedure is not reproduced here. The article says what the hazard is, points at the manufacturer’s own document and at the Safety Data Sheet for the specific product, and explains why the detail is being withheld. A safety claim that cannot be sourced this way is cut rather than softened, and this holds even where cutting it makes an article less immediately useful.

Original testing is welcome when the full record is there: what was measured, with what equipment, under what conditions, and how the result was checked. A claim of having tested, measured or observed something without that record does not publish.

What is not published

• A claim of having tested, measured or observed something without the full record.
• A number presented as a measurement without a source.
• A citation that does not resolve, or that does not support the sentence beside it.
• A fabricated credential, title or implied expertise.
• Handling, dilution or disposal guidance without a current product-specific Safety Data Sheet reference.
• Padding: scene-setting openers, a conclusion that restates the introduction, or a section that exists to satisfy a shape rather than to answer a question.

Corrections

Documents get revised and figures go stale, so corrections are routine rather than embarrassing. How they are reported and what happens to them is on the corrections page.