3 ms·

I just hope they don't stop to offer the papers in PDF. Even when I'm on a computer, I still prefer to read PDFs.

creatonez 3y ago
There is a taste component to it of course, but the history of PDF shows that it's the wrong format for reading on a computer. It was originally meant to be the end result of a publishing process before printing, a layer that sits right between the publishing software and the postscript that gets sent to the printer. This makes the PDF format quite inflexible for reading on a computer, with it being impossible to properly zoom or adjust the reading experience. Unfortunately many institutions and businesses have ignored its limitation because PDF turned out to be an obvious-but-naive to put a 'sheets of paper' metaphor into a computer system, which in the 1990s appealed to tech illiterate folks doing bare-bones computerization of existing paper systems. So later we got complicated and error-prone tools for editing PDFs, and many random additions to the spec to allow for unusual use cases.
- impendia 3y ago
  > This makes the PDF format quite inflexible for reading on a computer, with it being impossible to properly zoom or adjust the reading experience. As an academic researcher, generally speaking I also prefer PDF, and the inflexibility and static nature is a feature, not a bug. I appreciate the fact that a paper will appear the same everywhere, that I can refer to "the top of page 7", etc. The exception is if I wanted to just skim a paper; in this case, I think I'd prefer HTML. I'm a huge fan of what arXiv is doing here. It effectively preserves the status quo, while adding an additional option on the side. The HTML option might prove a little bit useful for me, and it is likely to prove extremely useful for people with disabilities.
  - creatonez 3y ago
    > I appreciate the fact that a paper will appear the same everywhere, that I can refer to "the top of page 7", etc. There are many great solutions to this problem, including ones that don't require Javascript at all. This website (https://gwern.net/silk-road https://gwern.net/silk-road) presents a really good example -- every header and sub-header is a clickable anchor. If more granularity is needed, on newer articles most of the paragraphs start with an italicized margin note -- though for technical writing, paragraph anchors might be better. The page also pays careful attention to print CSS and has a 'reader mode' to convert all links to footnotes when printed. Some websites will also preserve the text you select in a URL anchor, but more often than not this is just cumbersome. It also has a greater risk of not surviving changes to the webpage.
    - gwern 3y ago
      Also intriguing as a solution, which is potentially much stabler across revisions (page numbers are unstable), is the text-anchor-fragment feature Chrome introduced a while back: https://developer.mozilla.org/en-US/docs/Web/Text_fragments https://developer.mozilla.org/en-US/docs/Web/Text_fragments It's actually hit ~88% of the market https://caniuse.com/mdn-html_elements_a_text_fragments https://caniuse.com/mdn-html_elements_a_text_fragments but unfortunately, Firefox remains a holdout* and that's my browser, so I don't use it (although maybe I should just install https://addons.mozilla.org/en-US/firefox/addon/link-to-text-fragment/ https://addons.mozilla.org/en-US/firefox/addon/link-to-text-... and try it out - my existing method of making new anchors for annotation purposes is cumbersome). * Firefox officially is positive on it but no sign of any movement on it: https://mozilla.github.io/standards-positions/#scroll-to-text-fragment https://mozilla.github.io/standards-positions/#scroll-to-tex... https://github.com/mozilla/standards-positions/issues/194 https://github.com/mozilla/standards-positions/issues/194 https://bugzilla.mozilla.org/show_bug.cgi?id=1753933 https://bugzilla.mozilla.org/show_bug.cgi?id=1753933 https://wicg.github.io/scroll-to-text-fragment/ https://wicg.github.io/scroll-to-text-fragment/
aragilar 3y ago
I know of no-one who provides only HTML to arxiv, it's either latex, or doc/odt, so the PDFs should always be there.

There is a taste component to it of course, but the history of PDF shows that it's the wrong format for reading on a computer. It was originally meant to be the end result of a publishing process before printing, a layer that sits right between the publishing software and the postscript that gets sent to the printer. This makes the PDF format quite inflexible for reading on a computer, with it being impossible to properly zoom or adjust the reading experience. Unfortunately many institutions and businesses have ignored its limitation because PDF turned out to be an obvious-but-naive to put a 'sheets of paper' metaphor into a computer system, which in the 1990s appealed to tech illiterate folks doing bare-bones computerization of existing paper systems. So later we got complicated and error-prone tools for editing PDFs, and many random additions to the spec to allow for unusual use cases.

impendia 3y ago
> This makes the PDF format quite inflexible for reading on a computer, with it being impossible to properly zoom or adjust the reading experience. As an academic researcher, generally speaking I also prefer PDF, and the inflexibility and static nature is a feature, not a bug. I appreciate the fact that a paper will appear the same everywhere, that I can refer to "the top of page 7", etc. The exception is if I wanted to just skim a paper; in this case, I think I'd prefer HTML. I'm a huge fan of what arXiv is doing here. It effectively preserves the status quo, while adding an additional option on the side. The HTML option might prove a little bit useful for me, and it is likely to prove extremely useful for people with disabilities.
- creatonez 3y ago
  > I appreciate the fact that a paper will appear the same everywhere, that I can refer to "the top of page 7", etc. There are many great solutions to this problem, including ones that don't require Javascript at all. This website (https://gwern.net/silk-road https://gwern.net/silk-road) presents a really good example -- every header and sub-header is a clickable anchor. If more granularity is needed, on newer articles most of the paragraphs start with an italicized margin note -- though for technical writing, paragraph anchors might be better. The page also pays careful attention to print CSS and has a 'reader mode' to convert all links to footnotes when printed. Some websites will also preserve the text you select in a URL anchor, but more often than not this is just cumbersome. It also has a greater risk of not surviving changes to the webpage.
  - gwern 3y ago
    Also intriguing as a solution, which is potentially much stabler across revisions (page numbers are unstable), is the text-anchor-fragment feature Chrome introduced a while back: https://developer.mozilla.org/en-US/docs/Web/Text_fragments https://developer.mozilla.org/en-US/docs/Web/Text_fragments It's actually hit ~88% of the market https://caniuse.com/mdn-html_elements_a_text_fragments https://caniuse.com/mdn-html_elements_a_text_fragments but unfortunately, Firefox remains a holdout* and that's my browser, so I don't use it (although maybe I should just install https://addons.mozilla.org/en-US/firefox/addon/link-to-text-fragment/ https://addons.mozilla.org/en-US/firefox/addon/link-to-text-... and try it out - my existing method of making new anchors for annotation purposes is cumbersome). * Firefox officially is positive on it but no sign of any movement on it: https://mozilla.github.io/standards-positions/#scroll-to-text-fragment https://mozilla.github.io/standards-positions/#scroll-to-tex... https://github.com/mozilla/standards-positions/issues/194 https://github.com/mozilla/standards-positions/issues/194 https://bugzilla.mozilla.org/show_bug.cgi?id=1753933 https://bugzilla.mozilla.org/show_bug.cgi?id=1753933 https://wicg.github.io/scroll-to-text-fragment/ https://wicg.github.io/scroll-to-text-fragment/

aragilar 3y ago

I know of no-one who provides only HTML to arxiv, it's either latex, or doc/odt, so the PDFs should always be there.