It isn’t crazy, Adobe has copyrighted PDFs.
They locked it down so hard. The only way to mess with it is with pirated software.
Comment on Total fabrication
CkrnkFrnchMn@lemmy.ca 4 days ago
Just as crazy as not being able to convert PDF “properly” to docx or other format…
Honytawk@discuss.tchncs.de 4 days ago
Laser@feddit.org 4 days ago
No, that’s not actually the issue. The PDF format is open.
The actual (non-)issue is that PDF is a format that is not meant to be converted or otherwise be edited. Let’s say you have a page of text with two columns. For a PDF file, the concept of columns doesn’t exist. It’s just a series of individual letters placed exactly in a specific location on a page (and other graphical elements I guess, but that’s not the current point). I think strictly speaking PDFs don’t have to use the space character, since PDF doesn’t have the concept of a word, just individually placed letters.
Readers already have to do some heuristics to allow copying etc. but how are you going to convert this into a word file? How would you detect that there are text columns? What is a page header / footer and what is just fancy repeating letters?
Note that as soon as you get into more modern PDF, there’s also stuff like accessibility features (big fun!). There is support for Universal 3D.
Nobody sane would try to write a proper editor for PDF. You only write that if you want to make money off of it because a corporation somewhere doesn’t understand or control the pipeline that created a PDF that they need changed.
ArsonButCute@lemmy.dbzer0.com 4 days ago
Onlyoffice (vibe coded trash [that works]) allows you to do this. You can also just straight up edit most PDFs I’ve run through it. The exception is if its a collection of images, which many PDFs unfortunately are.
I_Has_A_Hat@lemmy.world 4 days ago
If it works, then why is it vibe coded trash? Do you mean anything coded with AI is trash by default despite literally no one solving this problem manually for 30 fucking years?
Evil_Shrubbery@thelemmy.club 4 days ago
(There are other pdf editors, it’s not an unsolved problem, however as others pointed out pdf is a sort of print as a definition. Like saving a .txt as a .png. Whoever made a png/pdf did that purposely. Also you can even ocr any print/pdf for like 30 years.)
ArsonButCute@lemmy.dbzer0.com 4 days ago
I didn’t mean to say trash. That was muscle memory. Most of the vibe-coded projects I’ve used have been trash.
notabot@piefed.social 4 days ago
Code can work but still be trash. The important question is whether it’s actually maintainable. Can issues be fixed without creating more problems? Is it archiected to allow likely new features to be added without having to rebuild large parts of the code? Trash can certainly be produced by humans too, but LLMs don’t really take any of this into consideration, as they don’t think, so they have a very high chance of producing trash.
BenLeMan@lemmy.world 4 days ago
I.e. not crazy at all. We can’t turn strawberry jam into fully formed fruits, either.