pdf is a compiled format for typeset text, so you need a pdf compiler. i use latex + tectonic. pandoc is also a popular alternative. “converting for editing” is like decompiling a program, you’re not guaranteed to get the same thing back as was put in. i never do that, i recompile instead. if i need text from a pdf i use pdftotext and cross my fingers because the formatting ain’t coming back out. any program that does replicate formatting just does a best guess.
no, you can’t edit an existing pdf, the nonstandard form filling extension notwithstanding. you can extract as much information as possible from it and recreate it. that’s what “pdf editors” are doing. and since it’s not officially supported, any edit can screw the file up.
the reason you can’t just edit it is that pdf is basically a container for program code that runs on printers. so you can have text interspersed with formatting information, or text with non-existent characters approximated by vector images, or text that’s been rendered to a raster image and is not actually in the document. then you have the fact that pdf can embed specialized fonts, compressed files, security measures, and even internal programs. and it’s all offset-based in there so you need to modify the entire file structure in order to get it working again after adding text. what’s worse, since any file with a pdf document in it is a valid pdf document according to the spec, less reputable “pdf editors” can just embed whatever shit they want. it’s a common malware vector.
it’s much safer to re-build the document from source. if you don’t have the source, there are tools to extract just the textual content.
Ok, this definitely helps in understanding how PDF works. However, I really do edit PDFs regularly and have no problems with the edited ones. Already mentioned it ITT, PDF-Xchange lets me do so many things that listing them would sound like an advertisement. Editing the existing text tends to mess it up, that’s true, but it’s not crucial for me and all sorts of other actions work almost perfectly.
You’re imagining some very ideal circumstances for working with PDFs that have nothing to do with my own needs, so I can’t really make use of your advice. :/
I frequently download book and journal article PDFs, scan books myself, and upload them online. And ofc read them.
Editing the PDFs in my case includes e.g. adding the outline/bookmarks that allow for easier navigation, adding OCR, cropping, splitting and rearranging the pages when the scanned images aren’t ideal, removing watermarks…
that sounds like actual typesetting work! i’m very surprised that you don’t get access to the source. usually when uploading to a journal they want the latex source.
Banking is very PDF heavy, and many of these PDFs have a ton of logic baked into them. Some of the loan documents do literally all of the math for you so the loan officer just inputs the amount, term and APR and the PDF outputs a fully-filled loan document. Its pretty magical to see until you peek under the hood at the code and oh-my-god-what-the-hell-how-did-this-ever-work-in-the-first-place-this-must-be-purgatory
then convert it to a proper format until you’re ready. editing a pdf is like decompiling and editing an exe file.
What program do you use to convert PDFs, what format do you convert them into for editing?
pdf is a compiled format for typeset text, so you need a pdf compiler. i use latex + tectonic. pandoc is also a popular alternative. “converting for editing” is like decompiling a program, you’re not guaranteed to get the same thing back as was put in. i never do that, i recompile instead. if i need text from a pdf i use pdftotext and cross my fingers because the formatting ain’t coming back out. any program that does replicate formatting just does a best guess.
I’m not sure if I’m following you - a compiler can be used to edit an existing PDF?
no, you can’t edit an existing pdf, the nonstandard form filling extension notwithstanding. you can extract as much information as possible from it and recreate it. that’s what “pdf editors” are doing. and since it’s not officially supported, any edit can screw the file up.
the reason you can’t just edit it is that pdf is basically a container for program code that runs on printers. so you can have text interspersed with formatting information, or text with non-existent characters approximated by vector images, or text that’s been rendered to a raster image and is not actually in the document. then you have the fact that pdf can embed specialized fonts, compressed files, security measures, and even internal programs. and it’s all offset-based in there so you need to modify the entire file structure in order to get it working again after adding text. what’s worse, since any file with a pdf document in it is a valid pdf document according to the spec, less reputable “pdf editors” can just embed whatever shit they want. it’s a common malware vector.
it’s much safer to re-build the document from source. if you don’t have the source, there are tools to extract just the textual content.
Ok, this definitely helps in understanding how PDF works. However, I really do edit PDFs regularly and have no problems with the edited ones. Already mentioned it ITT, PDF-Xchange lets me do so many things that listing them would sound like an advertisement. Editing the existing text tends to mess it up, that’s true, but it’s not crucial for me and all sorts of other actions work almost perfectly.
You’re imagining some very ideal circumstances for working with PDFs that have nothing to do with my own needs, so I can’t really make use of your advice. :/
in what circumstance does pdf editing come up regularly?
I frequently download book and journal article PDFs, scan books myself, and upload them online. And ofc read them.
Editing the PDFs in my case includes e.g. adding the outline/bookmarks that allow for easier navigation, adding OCR, cropping, splitting and rearranging the pages when the scanned images aren’t ideal, removing watermarks…
that sounds like actual typesetting work! i’m very surprised that you don’t get access to the source. usually when uploading to a journal they want the latex source.
Banking is very PDF heavy, and many of these PDFs have a ton of logic baked into them. Some of the loan documents do literally all of the math for you so the loan officer just inputs the amount, term and APR and the PDF outputs a fully-filled loan document. Its pretty magical to see until you peek under the hood at the code and oh-my-god-what-the-hell-how-did-this-ever-work-in-the-first-place-this-must-be-purgatory
yeah fun fact that’s usually an embedded javascript runtime
yet another reason for it to die in a fire