Quarto makes it difficult to get PDF and HTML outputs with consistent formatting, and other limitations for manuscripts

[a-dna-n]

Hello.

Quarto has some limitations that were difficult to pick up in the first place and time-consuming to figure out. For the sake of other people considering Quarto for a manuscript, I think it would be good to revise the documentation and e.g. link any preview to a specific output format.

For the record, Markdown tables have annoying limitations (headers and data justified separately, adjustable row height, etc.) so I read this and thought it would be fine to add formatting in HTML/css:

If you are comfortable with HTML, or use a package that generates HTML, the table cell HTML attributes colspan and rowspan are a better option. (Source)

I suppose I read too much into that. Please correct me if I'm mistaken, but here's what I experienced:

• In HTML outputs, tables span 100% of the width of the document by default. This is not apparent (or completely obscured) in the main introduction to tables: every narrow table is shown alongside the corresponding code and/or is pasted as a png instead of HTML. I figured this out, eventually, but it was perplexing.
• I couldn't find any examples of auto-numbering of HTML tables. This is another basic unmet need that led me to wonder if anyone actually tried to write an HTML-based manuscript with Quarto. (Inside the div? Outside the div? Colon?)
• The handling of CSS is oriented toward significant complexity: conversion of some tags, but not all, scss, Bootstrap etc. You can tell Quarto to leave R or Python as is, but if it's HTML, that rules out any output format but HTML. I got the sense that Typst would be the answer. (I tried that too).
• I could not find any examples of disabling HTML processing. I tried it again and again, but I couldn't know what that feature is actually supposed to do.
• I read the suggestion of using Pandas styling to format HTML tables. It adds its own quirks (R-inspired double row of headers, etc.) and vocabulary, but Quarto could end up modifying the format downstream. It would be useful (for developers and users alike) to see examples of actual HTML generated by Python instead of Python code.
• Last but not least, I could not find a single example of the same Markdown file being used to generate PDF and HTML outputs with the same format. I suppose it's only suggested, but there are always differences, and they're consistent. Here is a concise example:

The translation of the CSS means your table in Typst should look similar to how it looks in HTML without any intervention on your part.

The example that follows consists of these two tables. They're in separate tabs so they can't be seen at the same time.

Thanks.

[mcanouil]

@a-dna-n

Important

Accessibility: To improve accessibility, please add alternative text to your screenshots. This helps all users, including those using screen readers, to understand the context of the images. A brief description can make a big difference! See Good Alt Text, Bad Alt Text — Making Your Content Perceivable.

Important

Accessibility: For better accessibility, please ensure link text is descriptive and meaningful rather than using generic terms such as "here". This helps all users understand the link's destination. See Writing meaningful link text (WCAG).