This is the guide's correction log. When we find something wrong we fix it and record it here, with the date, the old claim, the new one and the evidence. We do not quietly edit a page and say nothing. Two corrections are logged so far, both from August 2026.
| Corrections logged | 2. Both found during research in August 2026 |
|---|---|
| Reported by readers | None yet. The log has only just opened |
| Subject of both | Dates in the island's own festival calendar |
| Where the right dates show | The trip dates tool, which carries the correction on the record |
| Where an old date remains | Three visitor pages, named below, at the time of writing |
| How to report an error | WhatsApp 07308 669617 |
| Typical fix time | Not established. This log is too new to claim one |
| What we will not do | Change a published claim and say nothing about it |
The wrong date was carried with a "One source" confidence marker, which is the marker doing its job. One source was enough to publish and not enough to be right. The correction did not come from finding a better festival listing. It came from the sinking date, which is documented, and from the fact that this kind of memorial is held on the anniversary rather than a week early.
Where the old date still shows. The trip dates tool has 14 September and carries the correction on the record itself, so it contradicts the older pages out loud rather than silently. Three pages still print 7 September at the time of writing: the best time to visit page, the September month page and the season page in the For businesses section. All three read from one place in the site data, so it is a single fix, and it is listed as open work rather than described here as done.
The old wording was not wrong so much as vague, and vague in a way that hides a real disagreement. "A Sunday, 2 to 8 July" describes a window and gives you nothing to plan around. The rule local sources actually give is the first Sunday of July, which is a rule a calendar can compute.
There is one year in seven where the sources genuinely conflict, and the guide now raises it only in those years rather than hedging every year. When 1 July is itself a Sunday, one reading gives 1 July and another gives the first weekend after 1 July, which would be 8 July. In those years the tool drops its confidence to low and says to check locally.
Where the old wording still shows. One page, the best time to visit page, still prints "A Sunday, 2 to 8 July" in its festival table. It is the same page that still prints 7 September for the Katsonis memorial, so both of this log's entries are visible as errors on it today.
This one is a useful example of what a correction is for. The visible change is small. The change underneath is from a range that could not be wrong to a rule that could be, which is the more honest position and the more useful one.
Message the guide on WhatsApp on 07308 669617. That is the only contact route on the site today, and pretending otherwise would be its own small dishonesty.
What helps, in order. The page you were looking at, ideally the address in the browser bar. The exact claim you think is wrong, quoted rather than described. What you believe the right answer is. And a source if you have one, which can be a photograph of a sign, an operator's own timetable, a municipal announcement or simply "I was standing there on Tuesday".
A source is not required. First-hand observation from somebody who was on the island is worth more than another listing site repeating the same figure, and it is exactly the kind of thing a desk-researched guide cannot get for itself. If you run a business here and something about your own place is wrong, say so and it gets fixed. That is not a favour and it does not buy anything.
Photographs are the same route. If you took one of the 110 images on this site and the credit is wrong, or you would rather it was not used, say so and it comes down.
Three outcomes, and all three end with an answer to you.
You are right and it is fixed. The page changes, an entry appears on this log with the date, the old claim, the new one and the evidence, and the old claim stays visible with a line through it. Deleting what a site used to say is how you make a correction log worthless.
You are right and it cannot be fixed yet. Sometimes the honest answer is that the sources now conflict and the guide should stop claiming either. In that case the number comes off the page and the page says the guide does not know. There are already three facts held in that state deliberately: the Skiathos to Skopelos crossing time, whether Agios Konstantinos is currently served, and whether any Skiathos bus has a wheelchair ramp.
We think the page is right. Then you get told why, with the source. If the disagreement is real and unresolved, that goes on the page as a stated conflict rather than being settled quietly in the guide's own favour.
What will not happen is a silent edit. If a published figure on this site changes, it should be possible to find out that it changed, when, and on what evidence.
The log is for factual errors: a wrong date, a wrong fare, a wrong stop number, a restaurant credited with something it does not do, a photograph of the wrong beach. Those are the things a reader can be misled by.
It is not for opinions. The five category scores on every restaurant page are editorial judgements and they are labelled as such. If you disagree with a score, that is a disagreement rather than an error, and the write-up already tries to give you enough to disagree with. Rewriting a score because a business asked would break the one promise this guide makes about money.
It is not for staleness either. Aggregate review scores drift daily and every one on this site is dated 16 August 2026. A rating that has moved since then has not become an error, it has become old, and the date on the page already says so. Timetables and fares work the same way and every page tells you to confirm them.
Typos and broken links get fixed without a log entry, because nobody was misled by a missing bracket.
The line, stated once. If a reader could have made a real decision on the strength of a claim, and the claim was wrong, it goes on this page with a date. If not, it just gets fixed.
Where the scores, the climate figures and the timetables come from, and what each of them cannot tell you.
What it covers, who it is for, and plainly how it is paid for.
Give it two dates and it says what the guide can honestly tell you about those exact days.