What this policy commits us to
- Every format is verified against the operator’s or standards body’s own specification — never against another generator site.
- The primary source is named on the page, so you can check us rather than trust us.
- Corrections are made on the page and the updated date moves with them. We never backdate.
- No invented authors, no invented credentials, no fabricated statistics.
- Nothing on this site is paid placement, because nobody pays us for anything.
This page describes how Tools SoShareIT decides what to publish and what has to be true before a value reaches a table. It exists because in reference data the failure modes are quiet: a wrong postal pattern looks exactly like a right one until somebody’s validation suite fails at three in the morning.
Last updated: 30 August 2026 · Published by the Tools SoShareIT editorial team
Why we refuse second-hand data
We do not take postal formats, ZIP ranges or reserved address blocks from other generator sites. Not as a starting point, not as a cross-check.
A wrong reference value is unusually durable. One site publishes it, three more copy it because it is already visible, and within a couple of years it is repeated everywhere without a single person having opened the specification that defines it. Nobody catches it, because catching it requires the exact step that was skipped. Copying is how that starts, so we do not copy.
What counts as a source
For a value to go on a page it has to come from the body that actually decides it.
| Kind of data | Accepted source | Not accepted |
|---|---|---|
| Postal code formats | The national postal operator; UPU addressing specifications | Other generator sites; crowd-edited lists |
| US ZIP ranges by state | USPS published assignments | Scraped datasets of unknown vintage |
| Reserved IP ranges | The RFC that reserves the block, cited by number | Blog posts summarising the RFC |
| Reserved domains | RFC 2606 and RFC 6761; IANA registries | Convention or habit |
| Test card numbers | The payment processor’s own documentation | Numbers that merely pass a Luhn check |
What we check before publishing a format
Four things, in this order. The order matters, because the first check changes what the other three are looking at.
Does the pattern match the current specification?
Run first, because a format that changed makes every downstream check meaningless. This is the one that catches South Korea — six digits until 2015, five digits since — and Ireland, which had no postcode system at all before the same year. A source that predates a reform looks perfectly authoritative and is simply wrong.
Does the generator produce something that matches?
We generate a batch and read them. A pattern documented correctly on the page and implemented incorrectly in the tool is a worse failure than either alone, because the page vouches for the bug.
Is the variability real or an artefact?
Some formats are genuinely variable — a UK postcode may be six, seven or eight characters, and M1 1AE and EC1A 1BB are both correct. Others only appear variable because a source was sloppy. Getting this wrong in the cautious direction produces a validator that rejects a third of real users.
Would this value collide with something real?
Applied to IP ranges, domains and card numbers rather than postal codes. A value goes on the reserved list only if a standards body has explicitly set it aside. “Probably unused” is not a category we publish, because the whole purpose of the table is that a reader can hard-code the value without thinking about it.
What we say about accuracy, and what we don’t
Every generated value on this site is format-valid and none of it is delivery-valid. We say so on the front page, in the terms, and next to the tool, because a reader who mistakes one for the other is the failure this site could most plausibly cause.
We do not claim our tables are complete, and we do not claim a generated code is unassigned. We do not publish an accuracy percentage, because we have no honest way to measure one and a made-up figure is worse than none.
Corrections
When we get something wrong, we change it on the page where it was wrong. We do not delete the page, we do not quietly edit and leave the date untouched, and we do not backdate to make the site look more consistently right than it was.
The last-updated date under each heading moves only when the content genuinely changed. That is what makes it worth reading: a date that moves on every automated touch tells you nothing at all.
Reader corrections are the most useful mail we receive and they are treated as such — the contact page says what to include so we can act on one quickly.
Independence
There is no advertising on this site, no affiliate link, no sponsored row and no paid inclusion. Nobody pays for a country to be added or for a tool to be recommended, so there is no commercial reason for a table to say anything other than what the specification says. We also run no analytics and set no cookies of our own, which the privacy policy covers in detail.
Authorship
Pages are published by the Tools SoShareIT editorial team and attributed that way. We do not invent a named expert, a photograph or a credential to make a page look more authoritative, because a fabricated byline is a lie about the most checkable thing on the page. Who we are is on the about page; how we work is here.