Hacker News Reader

Comment by aasasd | Hacker News Reader

People put in effort making PDFs and ensuring that nobody can read them without either a 14" portrait screen or a ton of scrolling—and you had to come along and ruin that carefully laid out inconvenience? What's wrong with you.

The incredibly complex layout of ‘a wall of text and occasional pictures’ is there for a reason. That's what the authors wanted, and only PDF is up to the task of representing such delicate formatting.

View on HN

Maybe not only PDF is up to the task, but this site doesn't show that web would be. In section "The quantum circuit model", look how 1/sqrt(2) looks (the 2 falls out of the sqrt sign). Then there is inline maths using the font "Noto Serif", which isn't suitable for writing formulas and makes it look really inconsistent when taken with the formulas which stand in their own lines. ℂ uses "Noto Serif" even in dedicated lines - the straight line with the rounded line couldn't be closer together. This font just isn't up to the task. It makes the formulas both less legible and ugly. Inconsistent spacing in "|0>" looks really off. A lot of inline formulas look as if they were about to fall off their lines. I don't get the use of sans-serif for section titles - default serif fonts in LaTeX make section titles both stand out and look like the belong in their surroundings --- this one doesn't have that effect. Hyphenation algorithm is also suboptimal. Look at the original PDF - at the same time there are fewer occurrences of hyphenation and the space between words in paragraphs is more evenly distributed. Also, the hyphens are easier to spot.

What PDF-s generated by LaTeX have, that so far web failed to replicate, is beautiful fonts for maths, consistency, legibility and overall beauty of the end result. They also have the advantage that they are self-contained, so if I want to archive one or move to another place, I just grab the single file and I'm done. With web a single document is distributed across many files and you need make a weird wget dance (which is easy to misuse)/install some obscure browser extensions/etc - not exactly user-friendly. Also, if you want to print the document to get away from the distraction of the computer, you're guaranteed to get at least as good a result as what you see on the screen.

Don't get me wrong. This attempt is better than any that I've seen so far. Kudos to the author for that. It's no small feat. But it doesn't prove that PDF-s are obsolete and that people favouring them are irrational.

Maybe you should look up MathML. Structured PDFs use MathML just like HTML to represent math formulae.

The real problem exists because most people don't use a correctly formatted/structured PDF to begin with. I don't wanna think about all the problems MS word might cause here and probably violates spec-wise.

Vectorized PDFs also don't use embedded fonts and content flow instructions but a bunch of randomly sequenzed glyphs that make no sense without a very good OCR. Vector-based PDFs are usually garbage for automated usage, and they are not useful anyhow for assistive purposes (e.g. a screenreader or a converter that uses the DAISY format or similar).

So yeah, I'd argue that PDF is the wrong serialization format. There are standardized alternatives that would be easier to parse, communicate, and license.

And countering your argument about portability: MHTML and WARC formats are very portable, the former is the default format for the page save functionality of all mobile smartphones. They are a single file, containing all necessary resources to display the page.

> This attempt is better than any that I've seen so far.

Have you seen arXiv Vanity? - https://www.arxiv-vanity.com

Not sure if this is sarcasm... but I don't think OP made this (https://github.com/dginev/ar5iv), they are just sharing.

Besides I don't think this is meant to be a total replacement that everyone is going to magically use and abandon PDFs, it might just be a convenience within a workflow to skim and check the full PDF if necessary.

I think the holes on your sarcasm filter might be too large.

I used to think I could recognize sarcasm whenever I saw it. Then I lived through the past few years.

I'm not making detailed arguments about what your sarcasm filter should be, but if this comment isn't detected as sarcasm, you should turn a knob somewhere.

I have to calibrate it to the fact that during the 2016 presidential debates one of the major party candidates assured everyone watching that his sex organs were doing great. And he won. My knobs are on 11, frankly.

I think a lot of people overestimate how well sarcasm can be determined on the internet. Sarcasm detection involves not only culture but also age, morals, and politics. Something that may appear as sarcasm to one person/group when addressed to an "in-group" would appear as someone actually saying that seriously to an "out-group", especially when it reinforces any kind of stereotype. Something that would be easily detected as sarcasm by a friend would be treated absolutely seriously by my parents, and my cousins who are almost a generation younger than me often write sarcasm I can't detect.

There's only one rule of sarcasm on the internet, if you want to be sarcastic, always end your comment with /s, otherwise avoid it.

To get back to the topic, I detected the OP's post as sarcasm, but I had to read it most of the way through it before I determined it was probably sarcasm, but I wasn't 100% sure.

I couldn't disagree more. The rule of sarcasm on the internet is the same as every other rule where everyone is anonymous. Give people the benefit of the doubt.

Moreover, the claim that sarcasm can't be detected on the internet is extremely dubious. It's really not that hard. The person puts extra effort into showing how the thing that they are pretending to believe is an absurd thing to believe. This is not culture or age dependent.

"It's awful how those people are doing all that good stuff" is a statement that requires you to know zero context. An alien can detect that this sentence is absurd. Moreover, if someone actually believes something absurd, and presents it this way, it doesn't matter if it's sarcasm or not, since they aren't convincing anyone.

The rules of communication trump the rule of sarcasm. If you want to get your message across clearly, avoid disambiguation. When the target audience misunderstands the message, that's on the author.

If I wanted to hunt for a hidden meaning in written text, I'd rather read poetry.

> It's really not that hard.

Correction: It's really not that hard for you. It IS hard for me, especially in recent years.

PDF is also bad for machine translation. It's a accessibility issue.

Reply on news.ycombinator.com