Notes from building this
Each of these came out of a decision made while building the tools here, rather than from reading about the subject. They are written for people who want to know why, not just what.
4 September 2026
Two ways to edit a PDF, and why one of them wrecks the document
Editors either draw on top of your page or rebuild it from scratch. The second is why your fonts change and your text stops being selectable.
2 September 2026
The photograph in that PDF may be bigger than the page shows
Placing an image in a document and cropping it usually stores the whole image and shows part of it. The rest is still in the file.
2 September 2026
Resize first, then compress: what actually makes a photo smaller
Most people reach for the quality slider and leave the dimensions alone. That is the wrong order, and it costs visible quality for very little saving.
1 September 2026
Why you cannot merge two PDFs by sticking them together
Both documents contain an object numbered 4, and they mean different things. Everything difficult about merging follows from that.
1 September 2026
WCAG 2.2 in practice: the rule that will fail your site
Of the nine criteria added in WCAG 2.2, one can be measured automatically and most sites fail it. It is about the size of the things you tap.
30 August 2026
What is actually inside the ZIP you are about to send
Archives are assembled by dragging a folder. Whatever was in that folder goes too, including the files your operating system put there without telling you.
27 August 2026
The safest way to edit a PDF is to move nothing
PDFs are held together by a table of byte offsets. Rewriting the file to remove a name is how you end up with a document that will not open.
24 August 2026
Why most metadata tools quietly fail on iPhone photos
HEIC does not store metadata where JPEG stores it. Tools written for JPEG find nothing and report nothing, which looks identical to a clean file.