X-Labs · Methodology
How research here is done, and how it is allowed to fail
The working rules behind every X-Labs article: primary sources with retrieval dates, blocked sources recorded as blocked, and a negative result published as a result.
On this page
- Written by
- Aldin Smajkan
- Type
- Methodology
- Published
- 15 September 2026
- Last revised
- 16 September 2026
- Version
- 4
What this answers: how a claim on this site is allowed to be made, and how a piece of work here is allowed to fail. Every other page is held to it. If you want the tool rather than the rules, start at Ask in your own words.
These are the rules the work follows. They are written down because a method that lives in somebody's head cannot be checked, and an article whose method cannot be checked is an opinion with footnotes.
A source is something you fetched, on a date
Every factual claim points to a source that was actually retrieved. Every source carries the date it was retrieved on.
That date is not bookkeeping. For legal material it is part of the claim. A text read in March and a text read in September may not be the same text, and nothing tells you which one you are holding. An undated citation quietly says the source is timeless. Almost none of them are.
Where a primary source exists, it gets used. A summary of a source is not a source. A search engine's description of a page is not a source. If the page could not be fetched, it was not read, and being confident about it does not close the gap.
A blocked source is written down as blocked
Sometimes a URL returns 403. Sometimes 404. Sometimes it serves an empty page because the content is drawn by JavaScript. When that happens we record the exact URL, the exact failure and the date.
We do not fill the gap in from memory.
This is the easiest place for a research document to go wrong, because the gap is invisible in the finished text. A confident paragraph about a page nobody opened reads exactly like a confident paragraph about a page somebody did.
So the survey of Bosnian legal sources contains a table of failures, and that table is one of the more useful parts of it.
Two kinds of time, kept apart
Anything about a rule needs two dates that move independently:
- When the rule was in force, or applied, in the world.
- When we retrieved and recorded it.
They are not the same date. Collapsing them is how a system ends up claiming that today's text was always the text.
A negative result is a result
A project here can end with "this does not work". That ending gets published with the same care as any other.
The failure it guards against is well known. A line of work that has already absorbed effort quietly narrows its claim until something survives. The defence is to write down in advance what would make the work stop, and then to publish it when it does.
What is never claimed
No article here claims peer review, institutional backing, or an experiment that was not run. Where somebody's involvement is possible rather than agreed, it is described as possible and they are not named.
Corrections
A corrected article keeps its address and its identity, gains a version number, and carries a dated line saying what changed. It is not silently edited and it is not replaced by a new article at a new URL.
A withdrawn article also keeps its address. It says it was withdrawn and why. An article that vanishes leaves every link to it lying, which is worse than the mistake it was hiding.
Why bother, on this site
X-Protocols argues that a claim is worth more when you can see its source and the time it was true. Research published under the same roof either meets that standard or undermines the argument.
In one line
Every claim here names a document, a date it was fetched, and what it did not say. Anything that cannot carry all three is written down as missing rather than filled in.
Sources
Every source below was retrieved on the date shown. A source read in March and the same source read in September are not necessarily the same document.
- Jedinstvena pravila za izradu pravnih propisa u institucijama BiH, Službeni glasnik BiH 11/05Retrieved 15 September 2026Copy hosted by the Indirect Taxation Authority, not by the gazette
Corrections
This article was changed after publication. It keeps its address and its identity; what follows is what changed and when.
- 16 September 2026Added an orientation line at the top naming the question this article answers, and pointed it at the question box at /labs/ask, which did not exist when this was written. No claim, source, date or commitment was changed, added or removed.
- 15 September 2026Added a closing line stating the rule these pages follow in one sentence. No rule was changed, added, relaxed or dropped.
- 15 September 2026Rewritten in plainer English. Shorter sentences, plain words, the point first. No rule was changed, added, relaxed or dropped, and the source list is unchanged.