Hacker News Reader

Comment by pyre | Hacker News Reader

PDFs are still ok though. There are many implementations of PDF parsers, many Open Source. It may not be as "universal" as plain-text, but it is definitely universal. At this point, I believe that all major operating systems ship with the ability to open PDFs for display.

View on HN

PDFs presume that the writer has "control" rather than the reader. Actually I probably shouldn't put quotes around "control" -- formatting is rigidly constrained, on purpose.

In addition PDF adapts, by design, a model of paper to the web. It's a "horseless carriage" file format.

Font size, page width, cut/paste and presentation in general should be the reader's choice, not the writer's. The Web manages this, sort of.

The OP is right on in this regard. Even the TeXs of this world, while better than the binary formats, have upgrade complexity.

>should be the reader's choice, not the writer's

Sometimes. There are a lot more design options available if the creator of the content maintains control over layout, fonts, etc. Sometimes this doesn't matter--if it's a block of text for example. Fairly simple layouts also render pretty well on the web.

Different content works better or worse with different approaches. One isn't intrinsically superior.

On the other hand, how does one write a standard invoice or a tax form using only plain-text / markdown? Formats with control over the layout have their place.

Maybe not just writer/reader, but also a presenter. In the context of education, a teacher will often be using someone else's material but presenting in their own style.

"PDFs are still ok though."

They are OK in theory, for the reasons you state.

The problem is the gratuitous use of PDF which we have all experienced - here's a common (pathological) example:

Document author starts with plain text - no special formatting or fonts, no images, etc. Somehow their toolchain converts that into a PDF file that contains No text, but rather an image of text.

The result is a big, bloated, unnecessary use of PDF that cannot even be parsed or used with anything but a graphical PDF viewer since the text is now gone - there is nothing but a picture. Of text.

A PDF without text still kind of works. It not accessible which is a big problem. However, this happens when there is a mistake somewhere in the chain. This is all too common, but what is the correct way to handle tables, diagrams, formulas, images, graphs, links in plain text. There is no right way to do these things in plain text. You will eventually need html or pdf for education.

Also: PDFs have pages and thus page numbers, which are an important tool. When working with long texts you need reference points to share your state.

That's what hyperlinks are. They even have the better property that they don't break if you change the font size for the whole text.

They don't mean much in a classroom of say 30 or 40. A lecturer will have n people with n texts and will have to synchronise them. In a typical lesson you jump to different pages also non-sequentially. The best option is a PDF etc. that you can optionally print. Anything else is a recipe for chaos, especially if you're not in a CS classroom.

Are they? I can't even copy text from a PDF without the new lines and indentation getting completely messed up never mind the rest of the formatting...

PDFs are awful for the blind, most PDFs are very painful to extract text from.

    PDFs are awful.

Could have stopped right there.

I agree that PDF is effectively universal as a distribution format for graphical and formatted text. There are some cases where plain text is still nicer, like when trying to read documentation without an X server.

PDF is a terminal format.

Even Pandoc gives up on trying to read it. To be fair, though, I think that the office suite file formats are even worse.

I have done a lot of work with code for handling pdfs.

The spec is 1000+ pages, references other docs, has many omissions and contains much that is apocryphal, or at least wildly inaccurate

pshaw.

right now, today, tuesday june 13th, 2017, safari will not open a .pdf to a specific page or bookmark, neither on the desktop nor in ios.

chrome can. firefox can. but safari cannot. which means neither the iphone nor the ipad can do it.

imagine if -- on the web -- you couldn't deep-link to an anchor in the middle of a webpage, but merely to (the top of) the webpage itself.

that's not the only deficiency of the .pdf format. it's not even the most galling one. it's just the one that happens to be hamstringing a certain project of mine at the moment. and it's illustrative.

PDFs seem to render poorly on ebook readers.

ebook readers yes. But they can work quite well on full-size tablets.

Kindles etc. are great when you're mostly reading a flow of text. For anything that benefits from design layout -- positioned graphics, sidebars, footnotes, etc. PDF on a 10" tablet is often better.

Ideally you'd provide a PDF as the most convenient format, but have a latex or similar root file that could be processed into other formats, like maybe .mobi or .epub.

Mainly because the layout code (Postscript?) in the PDF file assumes an A4 or US letter sized page. Most PDFs just cannot handle being reflowed.

> Mainly because the layout code (Postscript?) in the PDF file assumes an A4 or US letter sized page

Technically, it could be anything. The layout is specified, and everything stems from this.

Reply on news.ycombinator.com