Editorial standards

Every fact traces to a source we have read

A factual claim about the No One Left Behind pilots goes on a page only if it comes from a document we have read in full. The main one is the published evaluation of the Create@School approach (Gaeta et al.,Sensors 2019), open access under CC BY 4.0, obtained as full text. Before drafting began, its usable facts were extracted into a written source file: who took part, what was measured, what the authors report went wrong. Pages are written against that extract, not from memory: an extract can be checked line by line, a recollection cannot. An early draft of this site named an organisation as a project partner that was not one; it was caught before publication, and this practice is the consequence.

Where a source is out of reach we say so instead of paraphrasing an abstract. Some pages here were written before we had the full text, and say so on their face, naming the study without reporting its findings. Those pages are weaker than what we would write today, and better than a confident summary of a paper nobody here had opened. What we publish about the trial is atthe Spanish pilot schools, with the citation, the DOI and a link to the full text.

When two sources disagree, both get named

CORDIS puts the planned scale at some 600 children in five schools across three countries. The published evaluation describes the same experiment as reaching over 600 students from eight schools. Five and eight, one project, two sources each with a claim on being right, so both are printed and attributed. The same applies to scope: the paper reports the Spanish pilot only, three evaluation studies were planned, and we have not located the British or the Austrian one. Terms whose meaning a loose restatement would change are defined once inthe glossary.

Evidence and judgement are different sentences

Two kinds of sentence run through this site, marked differently. The first reports what a source establishes and is attributed: "the Spanish pilot's evaluation reports that". The second is what a practitioner would tell you, in our own voice and with a visible seam: "the evaluation does not speak to this, but in a classroom you would want to check".

The editor's background is in pedagogy and in teaching computing, and that is where the second kind comes from: what happens in the last ten minutes of a period, why a blank project is a poor place to start twenty-five pupils. It is deliberately not presented as research. No sample, no control group, no measurement, and where it sits beside a published finding the page says so.

What we do not do

We run no studies of our own, so nothing here is an effect size and no page claims that a tool caused anyone to learn more. The evaluation we lean on measures acceptance and user experience, and its authors close by saying that programming principles promoted logical thinking and stimulated pupils' creativity. That is their characterisation of what they saw, not a measured learning outcome, and rewriting it as proof that pupils learned more is the easiest wrong sentence available here.

We have not tested most of the tools we write about on shared school hardware with a full class on a school network. Where we have not tested something the page says so:the tool comparison puts its criteria and untested gaps at the top, and names no overall winner.

Nobody pays to appear here: no payment for coverage, no affiliate commission, no advertising, no sponsored entries, no link given in exchange for anything. The site runs no analytics and loads no third-party scripts; even the two typefaces are served from this domain. A page that carries no findings is sometimes the correct output:the publications page records what we have located and what we could not.

Corrections

If something here is wrong, we change it and say what changed. Not a silent edit: the page keeps a line naming what it previously said and when. There is nothing to show under it yet: the site is new, and nothing has been up long enough to be corrected in public. Changes to a draft that was never live are not corrections and will not be dressed up as a record of self-scrutiny. The route is incomplete too, because telling us that something is wrong needs an address and the contact page is not published, so this is for now an intention rather than a working mechanism.

Two languages, one set of facts

Every substantial page exists in English and in German with the same substance in both, and neither is a machine translation of the other. Where a literal rendering would produce something no German teacher says, we use the phrase they do: a mixed-ability class becomes "eine Klasse mit sehr unterschiedlichem Leistungsstand". Where one language needs a sentence the other does not, for instance that the Austrian evaluation is one we cannot find, it gets written.

What this set is still missing

The legal notice, the privacy statement, the accessibility statement and a contact address are not published. They exist as visibly unfinished pages kept out of search results, because unverified boilerplate would be a worse answer than an admission. Those four are ordinary obligations, not extras, and a site that asks to be trusted with claims about research does not get to treat them as optional. Until they are up, this page is the whole of the account.