💾 Archived View for yujiri.xyz › software › pdf.gmi captured on 2022-06-11 at 20:52:53. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2022-06-03)

➡️ Next capture (2023-01-29)

-=-=-=-=-=-=-

yujiri.xyz

Software

PDF must die

PDF is a blight and if you distribute PDFs you are making the world a worse place and more importantly, making me angry.

Some of the problems with PDF are symptomatic of a stubborn archaism, an attempt to bring the quirks and limitations of physical paper into the digital world:

• The concept of pages. Digital things are *scrollable* so pages are just pointless seams in the scrolling, often at egregious places such as between a heading and its first paragraph.

• The paper sizes PDFs can have: A4, Letter, Legal, etc.

Now, you might say these things actually do make sense for the rare use cases of printing to physical paper (the vast majority of such cases are themselves stubborn archaism). But if that's your only use, you wouldn't want to *distribute* documents in this format (as is often done), you'd want to only convert them to PDF as part of the printing process.

Some of the problems are shared with the web:

• Font is controlled by the document instead of by the viewer.

• PDF is just overall insanely complex. Ever tried measuring the source code of a PDF viewer?

• PDFs can apparently contain Javascript. Talk about an ill-fitting feature!

The Web is Evil

Other problems:

• Format is binary. You need dedicated software to read and edit PDFs. As such, you likely won't have access to advanced capabilities like searching by regular expression and extracting text for external transformation. This also means PDFs don't work with `diff`.

• Format is *proprietary*. This is super damning, come on. (And no, that you are allowed to implement it doesn't make it an "open standard"; that is a corruption of language invented by corrupt people.)

• Has a built-in password protection feature. This is silly. If you want to send a file that only a certain person can read, you either send it over a secure channel or encrypt the file itself, which works on any type of file. There is no need for such a feature at the file format level.

• PDFs can contain fillable forms!... but software support for this is spotty. Actually this is also symptomatic of a stubborn archaism. If you want information from someone, ask them for the information. No reason you need them to provide it by returning an edited copy of the document, and if you really want them to return an edited copy of the document, they can return an edited copy of the document (so you don't need form fields).

• Version issues. There are many different versions of the PDF format that you run into in the wild. As a software developer, I've had several experiences of errors caused by version differences (or at least error messages claiming this, as a red herring). Whenever I deploy code that deals with PDFs I have to be haunted by the possibility that a customer will come along with a PDF that doesn't work with my code and complain to us.

Replacements

Honestly with half the PDFs I run into, there's no real reason they couldn't have been plain text files. Most of the remainder would be better off as HTML files - they reside on the web anyway so being a PDF is just a pointless bit of friction (or a lot of friction if the user's browser doesn't display PDFs automatically).

A common case I see PDFs is legal shit. Government needs to be abolished and even besides any consideration of ethics virtually everything about all those institutions is stupid and broken and should be completely rethought.

Why you should be an anarchist

contact

subscribe via RSS