Converting a signed UDF to PDF: what happens to the signature

UDF is a document format by name, but technically it is an ordinary ZIP archive. Open it up and you see files. All of the text and formatting sits in a single file called "content.xml"; if the document has been signed, a separate file carrying the signature data sits beside it.

It helps to make that concrete. A real signed reply filing of a few pages takes up only 6.3 KB on disk. The archive holds two entries: a text file that unpacks to 17 KB, and a 2.6 KB signature file. The signature, in other words, is not something mixed into the appearance of the document; it is a separate record standing on its own inside the archive.

That structure has another consequence: a UDF made of text alone stays very small. A filing running into a size limit is not really possible; what runs into the limit is almost always the attachments, which is to say scanned documents. If you have a size problem, the place to look is not the filing but what is attached to it.

The tool that converts UDF to PDF opens the archive and reads only the text file. Whatever else the archive holds — the signature file included — is never touched. This is not an oversight but a design: what gets written into the PDF is the text and formatting of the document, while signature data is a UDF-specific record with no counterpart in PDF.

The conclusion is plain: a converted PDF is an unsigned document. A signature image or a signature ribbon showing on the page does not change that; what you see is text written into the document, not a signature backed by a certificate. For a PDF to count as signed, the signature has to be embedded in the PDF's own signature structure, and this conversion does not do that.

The same distinction applies to Dosya Kalemi's Sign PDF tool. It places a visible signature on the page — the digital counterpart of signing paper. It is not a qualified electronic signature, carries no certificate and does not prove who signed. Combining the two tools does not produce a signed document.

Where it really earns its place is when the person on the other end has no UYAP Editor. The editor needs a 32-bit Java installation to run, and on a 64-bit system setting it up often means editing environment variables by hand — UYAP's own help page walks through that step. When you are sending a document to a client, to an expert witness or to a colleague who does not use the editor, sparing them that installation is reason enough on its own.

The second place is reading and archiving. A PDF opens on every device, reads on a phone, files into a folder, and can be converted to a long-term archive format. Because the output is text-based, it can also be searched.

If the editor is already installed you may not need this tool at all: UYAP Editor converts UDF files in bulk to doc, odt, pdf and rtf through its batch save-as function. What this tool covers is the other side without the editor, and the machine where the installation cannot be done. The file staying put is part of that: the document is opened in your tab and written to PDF in your tab.

The working rule is this: the signed UDF is the original, and the PDF is a reading and archiving copy. Drawing that line up front is what keeps you out of the "I thought this one was signed" situation later.

In daily work this shows up in how files are kept. The signed UDF should not be deleted, should stay where it is, and its name should make clear which copy it is. When two versions of the same document sit in one folder under similar names, telling the original from the copy months later becomes hard.

It is also worth remembering that the conversion runs one way. There is no route back from PDF to UDF; if all you are left with is the PDF, the signed version cannot be reproduced. Converting should therefore be thought of as producing an extra copy alongside the original, not as a step that replaces it.

UDF is a format whose layout is not worked out in advance: the file holds text, formatting marks and paragraph properties, but nowhere does it say where a line ends. Whichever program opens the document works that out. So this tool is not a converter but a small typesetting engine: wrapping, alignment, indentation, list markers, page breaks and table layout are all computed from scratch.

Pixel-for-pixel sameness is out of reach for that reason, and so is the typeface: the font the editor uses cannot be shipped with a web page, so the output uses a serif face with close metrics. Line breaks landing a word or two apart is the natural consequence of that difference.

What is preserved is what makes the document recognisable: paragraph layout and alignment, bold, italic and underlined emphasis, list numbers and bullets, first-line indentation, headers and footers, tables and page breaks. The aim is a filing that reads and is recognisable, not a print-shop reproduction.