4 ms·

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 s

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.

fjorde 5y ago
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.
periheli0n 5y ago
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.
leephillips 5y ago
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/ 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.]

fjorde 5y ago

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.

periheli0n 5y ago

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.

leephillips 5y ago

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/ 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.]