Journal/Field notes

The photo that closes a defect, and the one that does not

The painter sent a good photo of a painted ceiling. Three weeks later the water mark was back. An after photo proves the trade did something, not that the defect is gone, and a fix needs two people on the record, not one.

A two-colour risograph poster of two overlapping photo frames drawn as plan rectangles, a single orange circle marking a point on the second
WP
Will PrinmanCo-founder
12 October 2026·5 min read

The photo is of a ceiling. It is a good photo: well lit, square to the wall, the cornice sharp. It was sent by the painter at 4:52 on a Thursday with the message "12.04 done". It is attached to a defect about a water mark on the ensuite ceiling in 12.04, and the water mark is not in it.

Not because the painter hid it. Because he painted the ceiling, the mark was gone, and he photographed the result, which is what you asked for. Three weeks later the mark is back, because the membrane above it was never fixed, and the file says the defect was closed on the Thursday with a photo attached. Which it was.

The after photo proves the trade did something. Not that the defect is gone.

That is the opinion, and it took a few of these to arrive at it. A rectification photo is evidence of work. It is not, on its own, evidence of a fix, because the trade who took it is the one person in the whole chain with a reason to send you a photo of the part that looks good. Not dishonestly. Just naturally. You photograph what you did.

A fix is proven by two things, not one. The person who did the work shows it. Then a person who did not do the work looks at it, against the original, and decides. Keep those two actions apart, with a name on each, and you have a record. Merge them into one tick and you have a painter closing a plumbing defect.

The before photo does half the job

Most of the argument about whether 12.04 was fixed is really an argument about what was wrong with it, and that argument is settled or lost on the day the defect was raised.

A water mark photographed from the doorway, with no pin on it, is a ceiling. The same mark photographed from underneath, with a circle drawn around the edge of the stain and a second photo of the shower niche above it, is a membrane defect. The painter looking at the second one knows that painting over it is not the job, and the site manager looking at his after photo knows what to look for.

IssuesId keeps the before photo's original device timestamp where the file has one, and who took it, and lets you pin and circle the fault on the image itself. That markup travels with the photo into every report, so the question "which bit of the ceiling" is answered once, by the person standing under it, and never again.

Gate the claim on the photo, and the argument stops

The single most common dispute on a job is not about quality. It is "I told you it was done" against "nobody told me". A trade says it on a phone call, the site manager forgets, the trade is annoyed, the item sits.

The fix is small and I would not run a job without it now. The trade cannot say the work is complete without attaching a photo of it. In IssuesId the Mark as Complete button on a trade's assignment is disabled until at least one photo is attached, and it says so. No photo, no claim. The painter who photographed the ceiling was doing exactly this, and that is the point: he made a claim with evidence, and the evidence is now sitting on the record for someone else to test.

The trade also has two other buttons, and they matter. I'm on it, which says work has started without pretending it is finished, and which most trades used to signal by doing nothing. And Dispute, for "this is not mine" or "I cannot get access", which raises it with the project team and leaves the reason in the thread instead of letting the item quietly stall.

The person who closes it was not holding the brush

Here is the part most registers get wrong by design. A trade in IssuesId has no defect transitions at all. None. Marking an assignment complete moves the defect to ready to inspect, which is a state that means "somebody from the builder's side needs to go and look", and nothing else.

A site user, a building manager or a manager then opens the before photo and the after photo side by side and, if it warrants it, walks up to 12.04. If the ceiling is painted and the niche above is still wet, they do not close it. They say so in the defect's conversation, at a visibility the painter can see, and the item stays with the trade. If it is fixed, they move it to closed, and the transition carries their name, their role and the time.

That is the second person. The whole value of the record is that it is there and it is different from the first.

Where the contract wants a third claim, "we formally accept this", sign-off is recorded separately: requested, then signed off by a configured approver with a timestamp and a note if they want one. Revoking it later stays in the history rather than erasing the original, which is the right behaviour and the opposite of what a spreadsheet does when someone untypes a tick.

Nothing in between gets tidied

A record that proves a fix is only as good as the bits of it nobody can quietly fix later. A note edited after posting shows as edited, with a count. A deleted note leaves a tombstone: something was removed here, at this time, by this person. Every email the defect sent, and whether it was delivered or bounced, sits on the trail. Read top to bottom, the history of 12.04 says: raised Tuesday with a pin on the stain, issued Tuesday, marked complete Thursday with one photo, not closed Friday with a note about the niche, marked complete again the following Wednesday with two photos, closed Wednesday afternoon by the building manager.

That is a fix. The Thursday photo on its own was a ceiling.

What to ask of your own register

Can the person who did the work close the record? If yes, your closed count is a count of optimism.

Does the record require anything of the trade when they say it is done? A tick is a claim. A photo is a claim with evidence.

Can you tell, for any closed defect, who looked at it and when, separately from who fixed it?

And when the verifier rejects a fix, does the record show the rejection, or does it just go back to looking open?

The short version of all this is at how do you prove a defect was fixed. The other half of the evidence problem, proving the defect was there in the first place, is at what makes a defect record stand up in a dispute. Between them they are the two photos.

IssuesId