Home/Research papers from Markdown to PDF
Use cases
Turn a Markdown draft into a paper ready to submit
A paper written in Markdown already has headings, a data table and a bibliography waiting inside it. Add footnotes and equations with dollar signs, then export straight to PDF in the Scholar theme's justified, academic page.
Equations with dollar signs
Wrap a short expression in single dollar signs to keep it inline with the sentence, for example $n = 24$. Put a longer equation between double dollar signs on its own line and it renders centred and larger. A results section usually sets a formula apart from the prose this way. The editor renders both forms as MathML, so no separate math stylesheet has to load or print.
The dollar sign is the only delimiter the editor reads. An equation copied with a backslash before the parentheses, or a backslash before the square brackets, stays as plain text on the page. Swap those pairs for a plain dollar sign before pasting a formula in. The same building blocks that write a mean also write a fraction, a square root or a summation sign for a standard deviation.
Footnotes that read like citations
Mark a reference in the sentence with a bracketed label, then write the source on its own line further down the document. Every reference collects into a numbered list under a rule at the end of the page. They appear in the order each one was first cited, the way a print journal sets its notes apart from the body text.
A data table from a CSV block
A method or results section often comes down to a short table of conditions, counts and a couple of measurements. Write those rows as comma-separated values inside a fence and they render as a proper table, with the first row read as the header. A semicolon-separated block works the same way, and the editor tells the two apart from the first line.
Worked example
A short study on line length and reading speed, with a method table, a results section carrying both an inline and a display equation, and three footnote references.
# Line Length and Silent Reading Speed on Screens
## Abstract
Twenty-four participants read short passages on a laptop display under two
column widths. A narrow column was set at 45 characters per line and a wide
column at 95 characters per line. Each participant read both conditions in
a counterbalanced order, so the design is within-subjects rather than two
separate groups. Reading speed was higher for the narrow column, in the same
direction as earlier print-based line length studies.[^1]
## Method
Participants were undergraduate volunteers with normal or corrected vision,
recruited through a department mailing list. Each session used three short
passages per condition, drawn from general-audience science writing and
matched for reading difficulty using a standard readability formula.[^2]
Passage order was randomised per participant, and each passage appeared in
only one of the two conditions to avoid a repeated-passage advantage.
```csv
condition,line_length_chars,participants,passages
Narrow,45,24,3
Wide,95,24,3
```
Both rows report the same 24 participants because everyone in the study read
both column widths. Reading speed was logged automatically as the number of
words displayed divided by the time the passage stayed on screen, then
converted to words per minute.
## Results
Mean reading speed for the narrow column was $\bar{x} = 264$ words per
minute. The wide column produced a lower mean of 241 words per minute, a
difference of 23 words per minute in favour of the narrow setting.
Variability across participants is summarised by the sample standard
deviation:
$$
s = \sqrt{\frac{\sum_{i=1}^{n}(x_i - \bar{x})^2}{n - 1}}
$$
For the narrow column s works out to 31 words per minute; for the wide
column s works out to 34 words per minute. The two distributions overlap
substantially even though the means differ, which is expected with a sample
of this size.
## Discussion
The gain for the narrow column matches the direction reported for printed
text, though it is smaller than the effect size in some of the earlier print
studies.[^3] One explanation is that screen reading already involves more
scrolling and smaller type than a printed page. That may narrow the room
for a line-length effect to show up. A between-subjects replication with a
larger sample would help separate the line-length effect from any residual
order effect that a within-subjects design cannot fully rule out. A next
step would be to repeat the comparison at a fixed font size across several
device widths, rather than a single laptop display.
[^1]: Line Length and Reading Rate in Print, Journal of Applied Typography, 2004, 8(1), 22 to 31.
[^2]: Readability Formulas for Short-Form Science Writing, Discourse Processes Quarterly, 2015, 41(3), 88 to 104.
[^3]: Column Width and Comprehension on Paper and Screen, Visual Communication Research, 2019, 14(2), 112 to 129.
Line Length and Silent Reading Speed on Screens
Abstract
Twenty-four participants read short passages on a laptop display under two
column widths. A narrow column was set at 45 characters per line and a wide
column at 95 characters per line. Each participant read both conditions in
a counterbalanced order, so the design is within-subjects rather than two
separate groups. Reading speed was higher for the narrow column, in the same
direction as earlier print-based line length studies.
Method
Participants were undergraduate volunteers with normal or corrected vision,
recruited through a department mailing list. Each session used three short
passages per condition, drawn from general-audience science writing and
matched for reading difficulty using a standard readability formula.
Passage order was randomised per participant, and each passage appeared in
only one of the two conditions to avoid a repeated-passage advantage.
Both rows report the same 24 participants because everyone in the study read
both column widths. Reading speed was logged automatically as the number of
words displayed divided by the time the passage stayed on screen, then
converted to words per minute.
Results
Mean reading speed for the narrow column was words per
minute. The wide column produced a lower mean of 241 words per minute, a
difference of 23 words per minute in favour of the narrow setting.
Variability across participants is summarised by the sample standard
deviation:
For the narrow column s works out to 31 words per minute; for the wide
column s works out to 34 words per minute. The two distributions overlap
substantially even though the means differ, which is expected with a sample
of this size.
Discussion
The gain for the narrow column matches the direction reported for printed
text, though it is smaller than the effect size in some of the earlier print
studies. One explanation is that screen reading already involves more
scrolling and smaller type than a printed page. That may narrow the room
for a line-length effect to show up. A between-subjects replication with a
larger sample would help separate the line-length effect from any residual
order effect that a within-subjects design cannot fully rule out. A next
step would be to repeat the comparison at a fixed font size across several
device widths, rather than a single laptop display.
Rendered in the Scholar theme. Open it in the editor to change the theme or the text.
Scholar is set for slow, printed reading
Scholar sets body text in EB Garamond at 19px, justified and hyphenated the way a printed page breaks its lines. The title sits centred at the top of the page, and a deep red marks the accent colour. Rules and table borders sit in a warm grey against a faint cream ground, close to the paper colour a printed journal page often has. It is built for essays, annotated bibliographies, course notes and anything meant to be printed and read slowly.
Paper is the alternative worth trying if the page reads better without justified lines. It sets Lora headings over Source Serif 4 body text on a warm off-white ground, with a rule under the title. Footnotes, tables and both kinds of equation render the same way under either theme.
Printing for submission
Once the paper looks right in the preview, use Print or save as PDF on the preview header. It opens the browser's own print dialog, and choosing Save as PDF there produces the file. That single file is what a submission system or an email attachment needs, without a separate export step.
Questions
Why does an equation show up as plain text instead of rendering?
A dollar sign only starts an equation when a non-space character follows it right away, and the closing dollar sign must not be followed by a digit. A backslash before an opening parenthesis or bracket is not read as a delimiter at all, so equations written that way stay as plain text.
Do footnote markers have to be numbers?
You can write any short label inside the brackets. The editor still numbers the references in order. It lists them under a rule at the end of the document, in the order each one was first cited.
Can I turn a spreadsheet of study conditions into a table?
Paste the copied rows into the source pane. The editor wraps tab-separated rows in a tsv fence and comma-separated rows in a csv fence, with an Undo button if you wanted the raw paste instead.
Does printing to PDF keep the Scholar theme intact?
Print or save as PDF opens the browser print dialog on the styled page as shown. Scholar is a light theme, so nothing gets inverted the way a dark theme does when it prints.