QR codes: what they actually do, and the mistakes that break them
Most QR code problems come from one misunderstanding: whether the code contains your link, or contains somebody else's link that redirects to yours.
Guides to the document problems people actually run into — why a PDF is too big to email, what OCR really does, and how to get a conversion that does not need an hour of tidying afterwards.
Most QR code problems come from one misunderstanding: whether the code contains your link, or contains somebody else's link that redirects to yours.
The document feeder does not know which way up the paper went in. Fixing the result takes two minutes; rescanning takes twenty.
A spreadsheet is an unbounded grid. Turning it into pages means deciding where to cut, and if you do not decide, the software decides badly.
The most common image mistake on the web is uploading the file straight from the camera. It is also the easiest one to fix.
Its origins are legal, but the problem it solves belongs to anyone assembling one document out of many: giving every page a name that cannot be confused.
A light blur over a house number can sometimes be reversed. A heavy one cannot. The difference is whether the detail was destroyed or merely disguised.
A typical paper devotes a quarter of every page to blank margin. On a tablet that blankness is stealing the space your eyes needed.
Splitting is not the same as compressing, and reaching for the wrong one wastes time. The choice depends on why the file is a problem.
Saving a photo as PNG makes it enormous. Saving a logo as JPG makes it fuzzy. The formats are not interchangeable, and the reason is how each one throws data away.
Platforms do not display your image. They display a version of it that fits their layout, and the difference is where your text goes missing.
A photo posted from home can carry the coordinates of the room it was taken in. Most people have no idea the data is there, because nothing ever displays it.
The tools are genuinely good now. They still fail on the same three things every time, and knowing what those are saves you a lot of retouching.
PDFs do not really contain tables. They contain text positioned so that it looks like one. That distinction explains every frustration you have had copying data out.
OCR accuracy depends far more on the quality of the image you feed it than on the software you use. Most failed recognitions are failed scans.
The whole promise of the format is in its name. When a PDF does not look the same everywhere, the cause is almost always a font that never travelled with the file.
Most people sign PDFs by pasting a picture of their signature. That is usually fine, occasionally not, and the difference is worth ten minutes of your attention.
Every year an organisation publishes a document with names blacked out, and someone recovers them in seconds by selecting the text underneath. Here is why that keeps happening.
There are two numbering systems in every PDF and nothing forces them to agree. That single mismatch causes an enormous amount of confused cross-referencing.
The most expensive changes are the small ones nobody mentioned. Reading two versions side by side is exactly the task human attention is worst at.
PDFs store where each character sits, not what order it should be read in. Everything wrong with copying from a PDF follows from that one fact.
Watermarks are communication, not protection. They are extremely good at the first job and almost useless at the second, and confusing the two causes real problems.
A two hundred page bundle does not need two hundred sheets. A few decisions before you press print typically cut it by three quarters.
Clicking a box and having nothing happen is not a broken reader. It usually means the form has no fields at all, and the fix depends on which kind you have.
You cannot add detail that was never captured. Modern upscalers get impressively close by inventing it, which is fine until it matters that the detail is real.
The document you exported is not always the document you were looking at. Word makes several decisions during conversion, and a few of them can be embarrassing.
The page you see and the page that prints are two different documents. Knowing that explains every strange PDF you have ever produced from a browser.
Nobody wants twelve separate photographs of your receipts. Almost every form, claim and application asks for one PDF, and there is a right way to build it.
A PDF page has no resolution until you render it. That single fact is why the same export can come out crisp or blurry depending on one setting.
That string of hex characters next to a download link is not decoration. It is the only way to know the file you received is the file that was published.
Length beats complexity by an enormous margin, and most of the rules people were taught produced passwords that were hard to remember and easy to guess.
Most mail servers reject attachments over 10MB, and scanned documents blow past that easily. Here is what actually makes a PDF large and how to fix it.
Someone has asked you to submit a PDF/A and you are not sure what that means or whether your file already qualifies. Here is the practical answer.
You can read the scanned contract perfectly well. Your computer cannot. Here is what OCR does about that, and how to get results worth relying on.
Some PDFs convert to Word almost perfectly. Others come out a mess. The difference is predictable once you understand what conversion is actually doing.
Sending payslips, medical letters or contracts by email? Encryption takes thirty seconds. Doing it properly takes one more decision that most people get wrong.
A dozen receipts, four authors' sections, or a contract whose signature page came back separately. Here is how to combine them properly.
Phone photos are several megabytes each, and almost none of that is doing any visible work. Here is how to reclaim it.
Someone wants page 12 of your thirty-page agreement. There are three tools that could do this, and picking the right one saves a lot of clicking.
Before you assume the file is lost: most PDFs that refuse to open are recoverable, and the fix depends entirely on which of five things went wrong.
You are about to upload a contract to a free website. That instinct to hesitate is correct — here is what to check, and why some tools never upload at all.