Hacker News Reader

Comment by cprecioso | Hacker News Reader

Using Markdown with pandoc and pandoc-crossref is basically perfect as a substitute of LaTeX. It allows in-line and standalone LaTeX, and citations with crossref.

View on HN

Pandoc completely solves any of the issues mentioned about markdown for scientific writing.

it works flawlessly, and it unarguably way cleaner than writing in plain latex

How can anyone

- tell me

- that this

- list

Is harder to read:

\begin{itemize}

  \item than this

  \item unintuitive

  \item mess

\end{itemize}

With the former, can you items of many paragraphs and other complex stuffs one sometimes need in a complex/advanced document? So you can understand why LaTeX had to go the less naive way MarkDown did.

Of course the LaTeX syntax is more capable, but for the simple and common case, it's extremely fussy. The simple case should be effortles. With markdown it is, and in Pandoc you can easily drop to more complicated syntax should you need any of its capabilities.

Are there any tutorials you could recommend please?

The official manual is pretty complete: https://pandoc.org/MANUAL.html.

In my experience this has zero advantage over LaTeX since converting your markdown to formats other than LaTeX will be just as hard as converting from a LaTeX source. It's even harder because you have to deal with an eclectic mix of LaTeX and markdown, and various errors and restrictions in pandoc.

yes, these are all fraught. Any X-to-Y format exchange is a brew/apt miracle in my experience. It's not necessarily important for a LaTeX document to be render-able in all media, but the maths disciplines, for example, should be asking themselves how they want to be able to search for things including formula, algorithms etc, and finding ways to guarantee that a maxima of media support search/find/read operations on their articles. It's not just about how easy is it to input? but it's very similar to "how easy is it to get back out?" it has been a bit of a mismatch in maths re formulae expressions, and how those are represented visually vs. textually.

The problems in converting LaTeX usually come from third-party packages. When only a restricted set of packages is allowed then it is quite straightforward. This is the reason many publishers require latex, even if they don't use LaTeX internally (which is often the case).

They convert LaTeX to the XML dialect that their in-house layout system understands.

I use the system to write documents that I convert to PDF (through LaTeX), Word, plain text, and HTML, with equations, tables, and internal and external links preserved (citations in plain text turn into footnotes). It works brilliantly, and this would be impossible when starting with LaTeX source. You can embed the target formats directly into the markdown when needed, and have alternative versions for different targets. I do this for things like tables, where I want to directly format them for LaTeX and HTML, rather than rely on Pandoc’s automatic translation. You can write filters¹ that extend Pandoc/markdown to do any processing that you want, so you can create your own syntax with custom translations to any target format.

[1] https://lee-phillips.org/panflute-gnuplot/ [This is sort of obsolete now, as you should probably write filters with Lua. But it’s still a good illustration of the possibilities.]

Reply on news.ycombinator.com