Merge pull request #234 from SirWumpus/markdown_O1

Refactor markdown guidelines to improve space and presentation.
This commit is contained in:
Landon Curt Noll authored and GitHub committed 2025-11-26 23:02:34 -08:00
commit a039928179
1 file changed
+258 -600
+258 -600
View File
@@ -1,717 +1,375 @@
<div id="markdown_guidelines">
# IOCCC 28th Markdown Guidelines
</div>
<div id="markdown_version">
## IOCCC Markdown guidelines version
</div>
# IOCCC 29th Markdown Guidelines
<p class="leftbar">
These [IOCCC markdown guidelines](markdown.html) are version **28.4 2025-04-28**.
</p>
## Markdown Guidelines Version
**IMPORTANT**: Be sure to read the [IOCCC rules](next/rules.html) and [IOCCC
guidelines](next/guidelines.html).
These [Markdown Guidelines](markdown.html) are version **28.4 2025-04-28**.
<div id="introduction"></div>
Jump to: [top](#)
## Introduction
<div id="introduction">
## IOCCC Markdown Guidelines Introduction
</div>
The IOCCC makes extensive use of [markdown](https://daringfireball.net/projects/markdown/). For example, when
submitting to the IOCCC we have people submit remarks about their
entry in markdown format. Every [winning entry](years.html) uses a `README.md`
markdown file as the basis for forming the `index.html` web page for
that entry. All HTML 5 generated pages on the [Official IOCCC website](https://www.ioccc.org/index.html)
start with some markdown content. We no longer support older HTML
standards and some of our guidelines help integrate winning entries into
the website more easily.
The IOCCC makes extensive use of [markdown](https://daringfireball.net/projects/markdown/).
For example, when submitting to the IOCCC
we have people
submit remarks about their entry in markdown format. Every
[winning IOCCC entry](years.html) uses a `README.md` markdown file
as the basis for forming the `index.html` web page for that entry.
All generated HTML pages on the [Official IOCCC website](https://www.ioccc.org/index.html)
start with some markdown content.
See the [markdown syntax](https://www.markdownguide.org/basic-syntax) guide. See also the [CommonMark Spec](https://spec.commonmark.org/current/).
See the FAQ on "[remarks.md submission file](faq.html#remarks_md)".
See the [markdown syntax](https://www.markdownguide.org/basic-syntax) guide.
See also the [CommonMark Spec](https://spec.commonmark.org/current/).
See the
FAQ on "[remarks.md submission file](faq.html#remarks_md)".
Nevertheless, the IOCCC does have certain practices that we ask authors to
follow. Some of these relate to use of markdown directly and others relate
to injecting HTML into the markdown file.
Nevertheless, the IOCCC does have certain practices that we ask authors to follow.
Some of these relate to use of markdown directly and others relate to injecting HTML
into the markdown file.
In the sections below, examples will mark by ❌ for the incorrect method
and ✔️ for preferred correct one.
Jump to: [top](#)
<div id="sections"></div>
<div id="sections">
## Please DO start `remarks.md` sections at level 3 (`###`), increasing for subsections up to 6 (`######`)
</div>
## Start `remarks.md` sections at level 3
IF (it is not required) you do use sections in your `remarks.md` file (this does
not apply to other markdown files), **PLEASE** start at level three:
If (it is not required) you do use sections in your `remarks.md` file (this does
not apply to other markdown files), **PLEASE** start at level three up to level six:
``` <!---markdown-->
# Foo <=== no thank you
## Foo <=== no thank you
```<!---markdown-->
❌ # Section Reserved
❌ ## Section Reserved
✔️ ### Section Level 3
✔️ #### Section Level 4
✔️ ##### Section Level 5
✔️ ###### Section Level 6
```
Instead, start sections with 3 `#`s:
<div id="anchor-name"></div>
## Replace `<a name="anchor_name"></a>` construct.
``` <!---markdown-->
### Section 0
```<!---html-->
❌ <a name="old_way">...</a>
✔️ <span id="new_way">...</span>
```
For each subsection of the current section, add another `#`, up to 6, going back
to 3 for new sections. For instance:
<div id="links"></div>
## Markdown links instead of `<a>...</a>`
``` <!---markdown-->
### Section 0
...
#### 1st subsection of section 0
...
##### 1st sub-subsection of the 1st subsection of section 0
...
#### 2nd subsection of section 0
...
##### 1st sub-subsection of the 2nd subsection of section 0
...
##### 2nd sub-subsection of the 2nd subsection of section 0
...
###### 1st sub-sub-subsection of the 2nd sub-subsection of the 2nd subsection of section 0
...
### Section 1
...
### Section 2
...
Etc.
```<!---html-->
❌ This <a href="#link_target>HTML link text</a> works, but we prefer
✔️ [markdown link text](#link_target)
```
<div id="del"></div><div id="strike"></div>
<div id="name">
<div id="anchor-name">
## Please do NOT use the `name` attributes in HTML `<a>...</a>` hyperlink elements
</div>
</div>
## Replace `<strike>` or `<s>`
Please do **NOT** use the HTML construct:
```<!---html-->
❌ <strike>striked through text</strike>
❌ <s>HTML 5 striked through text</s>
``` <!---html-->
<a name="string">...</a> <=== no thank you
✔️ <del>HTML 5 striked through text</del>
✔️ ~~markdown striked through text~~
```
as those are **NOT** part of the HTML 5 standard.
<div id="ins"></div><div id="underline"></div>
Instead use:
## Replace `<u>`
``` <!---html-->
<div id="string">...</div>
```<!---html-->
❌ <u>old tags</u>
✔️ <ins>HTML 5 tags</ins>
```
where the `...` is the HTML you want to "name". The `<div>`
encapsulates the HTML you want to name: i.e., the target of some
`<a href="#string">` or some other `<a href="URL#string">`
for the given page URL.
<div id="tt"></div>
### IMPORTANT POINT:
## Replace `<tt>`
There are certain markdown constructs that **CANNOT** have an **internal** `<div
id="string">...</div>` element.
An example is headings (lines that start with a `#`). For example:
``` <!---markdown-->
# <div id="string">THIS WILL NOT WORK!</div> <=== this will not work
```<!---html-->
❌ <tt>obsolete HTML</tt>
✔️ <span style="font-family: monospace;">monospace font</span>
✔️ `markdown style monospace`
```
For things like headings, you have to surround them with the `<div
id="string">...</div>` element, as in:
<div id="indent"></div>
``` <!---markdown-->
<div id="string">
# This will work
</div>
## Indent code blocks by four spaces, not by TAB.
```<!---sh-->
❌ echo "This code block is NOT indented"
```
While some browsers will still recognize the HTML construct `<a
name="string">...</a>`, it is possible **they MIGHT NOT** in the future.
We request that you indent the code block by multiples of four (4)
ASCII SPACES, never TAB (␉):
Jump to: [top](#)
````<!---sh-->
✔️ echo "This code block is indented by multiples of 4 spaces"
<div id="links">
## If you can, it is PREFERABLE to use markdown links rather than `<a>...</a>`
</div>
It is easier and preferred to use markdown links rather than HTML `<a>..</a>`
anchors.
Instead of:
``` <!---html-->
Use of <a href="#links>HTML anchors</a>
is one option, however ...
```
``` <!---markdown-->
[markdown links](#links) are easier and preferred
```
Jump to: [top](#)
<div id="strike">
<div id="del">
## Please do NOT use the `<strike>` or the `<s>` HTML element
</div>
</div>
Please do **NOT** use the obsolete `<strike>` or the obsolete `<s>`
(<del>**strikeout**</del>) HTML elements:
``` <!---html-->
<strike>...</strike> <=== no thank you
<s>...</s> <=== no thank you
```
Use instead:
``` <!---html-->
<del>...</del>
```
Jump to: [top](#)
<div id="underline">
<div id="ins">
## Please do NOT use the `<u>` HTML element
</div>
</div>
Please do **NOT** use the obsolete `<u>` (<ins>_underline_</ins>) HTML element:
``` <!---html-->
<u>...</u> <=== no thank you
```
Use instead:
``` <!---html-->
<ins>...</ins>
```
Jump to: [top](#)
<div id="tt">
<div id="span">
## Please do NOT use the `<tt>` HTML element
</div>
</div>
Please do **NOT** use the obsolete `<tt>`
(<span style="font-family: monospace;">**teletype**</span>) HTML element:
``` <!---html-->
<tt>The tt element is obsolete</tt> <=== no thank you
```
Instead use either a monospaced `<span>` or an inline markdown code block:
``` <!---html-->
<span style="font-family: monospace;">Use of a monospaced font
is one option,
however ... </span>
```
We recommend using the inline markdown code block method instead:
``` <!---markdown-->
Using the `inline markdown code block` is easier and is **preferred**.
```
Jump to: [top](#)
<div id="unindented">
<div id="indented">
## Please do NOT use unindented code blocks
</div>
</div>
Please do **NOT** start code blocks at the first column.
For example:
```` <!---markdown-->
``` <%%NO_COMMENT%%!---sh-->
echo "This code block is NOT indented" <=== no thank you
```
✔️ # The top level starts with a 4 ASCII space indent.
✔️ #
✔️ # Only indent with multiples of 4 ASCII spaces.
✔️ # 2nd level indent needs 8 ASCII spaces (no tabs).
✔️ # 3rd level indent needs 12 ASCII spaces (no tabs).
✔️ # etc.
````
We request that you indent the code block by multiples of 4 ASCII **SPACES**:
Alternatively its possible to use ` ``` ` marks, without the initial
four (4) space indentation. This is usually the better method.
```` <!---markdown-->
``` <%%NO_COMMENT%%!---sh-->
echo "This code block is indented by mutiples of 4 spaces"
# The top level starts with a 4 ASCII space indent.
#
# Only indent with multiples of 4 ASCII spaces.
# 2nd level indent needs 8 ASCII spaces (no tabs).
# 3rd level indent needs 12 ASCII spaces (no tabs).
# etc.
```
````<!---markdown-->
✔️ ```
✔️ The same thing applies to any markdown block bracketed by ``` lines.
✔️ ```
````
Moreover:
<div id="tabs"></div><div id="spaces"></div>
```` <!---markdown-->
```
The same thing applies to any markdown block surrounded by ``` lines.
```
````
## Do not use TAB in markdown files.
Please do **NOT** indent using ASCII tab characters in markdown files.
Please do **NOT** indent with ASCII TABs (␉), **OR** use any ASCII TAB
(␉) anywhere inside a markdown file.
Jump to: [top](#)
<div id="tabs">
<div id="spaces">
## Please do NOT use ASCII tabs in markdown files
</div>
</div>
Please do **NOT** indent with ASCII tabs, **OR** use any
ASCII tab characters anywhere inside a markdown file.
While we have nothing against the ASCII tab character in general,
we have discovered that ASCII tab characters create problems when
While we have nothing against the ASCII TAB (␉) character in general,
we have discovered that ASCII TAB (␉) characters create problems when
used as part of the leading whitespace within a markdown file.
If you need to indent 2 or more levels, use multiples of 4 ASCII
**SPACES**.
If you need to indent two (2) or more levels, use multiples of four (4)
ASCII SPACES.
For example:
```` <!---markdown-->
````<!---markdown-->
```
Please do **NOT** use ASCII tabs in markdown files. <=== no thank you
Please do **NOT** indent markdown with ASCII tabs. <=== no thank you
❌ ␉Please do **NOT**␉use ASCII tabs␉in markdown files.
❌ ␉␉Please do **NOT** indent markdown with ASCII tabs.
We don't mind tabs in general, <=== no thank you
We just ask that you NOT use them in markdown files.
❌ ␉We don't␉mind␉tabs␉in␉general,
❌ ␉We just ask that you NOT use them in markdown files.
```
````
And to clarify, we are only talking about markdown files,
not C code or any other non-markdown content:
And to clarify, we are only talking about markdown files, not C code or
any other non-markdown content:
```` <!---c-->
printf("It is fine to use tabs in Obfuscated C code.\n");
/* if you wish */
````<!---c-->
✔️ printf("It is fine␉to␉use tabs in Obfuscated C code.\n");
✔️ /*␉if␉you␉wish */
// We ask that you to NOT use ASCII tab characters in your remarks.md writeup,
// or in any other markdown files within your entry, please.
✔️ // We ask that you to NOT use ASCII tab characters in your remarks.md writeup,
✔️ // or in any other markdown files within your entry, please.
````
**NOTE**: Again, you are **perfectly welcome** to use ASCII tab characters in
your C code and other non-markdown files. We simply ask that you do **NOT** use any
ASCII tab characters in markdown files.
**NOTE**: Again, you are **perfectly welcome** to use ASCII tab characters
in your C code and other non-markdown files. We simply ask that
you do **NOT** use any ASCII TAB (␉) characters in markdown files.
Jump to: [top](#)
<div id="code"></div><div id="languages"></div>
<div id="vim-tabs">
### Tip for `vim` users
</div>
## Use HTML comments for code block language syntax.
If you use `vim` you can put in your `.vimrc` file (usually `~/.vimrc`) the
following settings to make sure the tabs are not put in without you noticing:
``` <!---vim-->
set tabstop=8 " a tab is 8 spaces (or whatever you wish it to be set to)
set softtabstop=4 " ...but when inserting/backspacing use 4 spaces
set shiftwidth=4 " ...and auto-indent 4 spaces (when autoindent is set)
set expandtab " ...expand tab into spaces
```
If you have file type detection enabled you can, if you prefer, have these
settings set just for markdown files:
``` <!---vim-->
autocmd! Filetype markdown setlocal set tabstop=8 softtabstop=4 shiftwidth=4 expandtab
```
or so.
This will prevent the tab key from inserting tabs; rather it will insert 4
spaces.
To **VERIFY** that there are no tabs in a file you may do, in command mode:
```
/\t
```
If you're in insert mode hit `ESC` first.
Jump to: [top](#)
<div id="languages">
<div id="code">
## Please use HTML comments to specify a language for a code block
</div>
</div>
We request that [fenced markdown code
blocks](https://www.markdownguide.org/extended-syntax/#fenced-code-blocks)
We request that [fenced markdown code blocks](https://www.markdownguide.org/extended-syntax/#fenced-code-blocks)
**NOT** specify a language directly.
For example:
```` <!---markdown-->
```c <=== no thank you
int main(void) {return 0;}
```
````<!---markdown-->
❌ ```c
❌ int main(void) {return 0;}
❌ ```
````
Instead, put the language inside an HTML comment, separated from the
markdown code block starting fence by a space:
```` <!---markdown-->
``` <%%NO_COMMENT%%!---c-->
int main(void) {return 0;}
```
````<!---markdown-->
✔️ ```<!---c-->
✔️ int main(void) {return 0;}
✔️ ```
````
**IMPORTANT**: The **initial** &nbsp; **\` \` \`** &nbsp; must be followed by an **ASCII SPACE**,
and **THEN** an **opening** **`<!---`** (a "`<`", a "`!`" and then three "`-`"s), and
**THEN** the **language** and **FINALLY** a **closing** "`-->`" (two "`-`"s
**IMPORTANT**: The **initial** &nbsp; **\` \` \`** &nbsp; must be
followed by an **ASCII SPACE**, and **THEN** an **opening** **
`<!---`** (a "`<`", a "`!`" and then three "`-`"s), and **THEN**
the **language** and **FINALLY** a **closing** "`-->`" (two "`-`"s
followed by a "`>`").
<div id="code-and-text"></div>
Jump to: [top](#)
## Ensure blank line between code block and text.
<div id="slash">
<div id="void">
## Please do NOT add trailing slash to void HTML elements
</div>
</div>
Please do **NOT** place text on the next line after a markdown code
block. Instead, place a blank line after the end of a markdown code
block as this makes it easier to detect when markdown code blocks are
**NOT** properly indented.
Please do **NOT** use a trailing slash on [void HTML
elements](https://github.com/validator/validator/wiki/Markup-»-Void-elements).
````<!---markdown-->
```
int
main(int foo)
{
return foo;
}
❌ ```
❌ C compilers cannot be given a -Wno-main-arg-errors flag.
See also this note on [trailing slashes in void-element start
tags](https://github.com/validator/validator/wiki/Markup-»-Void-elements#trailing-slashes-in-void-element-start-tags-do-NOT-mark-the-start-tags-as-self-closing).
The trailing slash on void HTML elements has no effect and interacts badly with
unquoted attribute values.
For example, please do **NOT** use markdown like these:
``` <!---html-->
<br/> <=== no thank you
<hr/> <=== no thank you
<img src="1984-anonymous-tattoo.jpg"
alt="image of a tattoo of the 1984 anonymous C code"
width=600 height=401 /> <=== no thank you
```
Instead use, respectively:
``` <!---html-->
<br>
<hr>
<img src="1984-anonymous-tattoo.jpg"
alt="image of a tattoo of the 1984 anonymous C code"
width=600 height=401>
```
Jump to: [top](#)
<div id="backslash">
<div id="br">
## Please do NOT use a TRAILING backslash (`\`) outside of a code block
</div>
</div>
Unless the line is inside a markdown code block, please do **NOT**
end a markdown line with a trailing backslash (`\`). Instead use
a trailing `<br>`.
Instead of:
``` <!---markdown-->
In markdown,\ <=== no thank you
do NOT use trailing\
backslashes outside of\
a code block
```
use:
``` <!---markdown-->
In markdown,<br>
use trailing<br>
br's outside of<br>
a code block
```
Again, use of a trailing backslash (`\`) inside a markdown code block is fine:
```` <!---markdown-->
```
This is OK\
inside a\
markdown code\
block
```
```
int
main(int foo)
{
return foo;
}
✔️ ```
✔️
✔️ C compilers cannot be given a -Wno-main-arg-errors flag.
````
as is:
<div id="tag-slash"></div>
``` <!---markdown-->
`This is OK\`
## Do not add trailing slash to void HTML tags.
Please do **NOT** use a trailing slash on
[void HTML elements](https://github.com/validator/validator/wiki/Markup-»-Void-elements).
See also this note on [trailing slashes in void-element start tags](https://github.com/validator/validator/wiki/Markup-»-Void-elements#trailing-slashes-in-void-element-start-tags-do-NOT-mark-the-start-tags-as-self-closing).
The trailing slash on void HTML elements has no effect and interacts
badly with unquoted attribute values.
```<!---html-->
❌ <br/>
❌ <hr/>
❌ <img src="1984-anonymous-tattoo.jpg" alt="image of a tattoo of
the 1984 anonymous C code" width=600 height=401 />
✔️ <br>
✔️ <hr>
✔️ <img src="1984-anonymous-tattoo.jpg" alt="image of a tattoo of
the 1984 anonymous C code" width=600 height=401>
```
Doing it this way will prevent `pandoc(1)` from generating deprecated HTML
elements such as `<br />`.
<div id="br"></div><div id="backslash"></div>
Jump to: [top](#)
## Do not use a trailing backslash outside of a code block
<div id="images">
<div id="img">
## Please do NOT use markdown style images
</div>
</div>
Unless the line is inside a markdown code block, please do **NOT** end
a markdown line with a trailing backslash (`\`). Instead use a trailing
`<br>`.
Please do **NOT** use the markdown embedded image element.
```<!---markdown-->
❌ In markdown,\
❌ do NOT use trailing\
❌ backslashes outside of\
❌ a code block
✔️ In markdown,<br>
✔️ use trailing<br>
✔️ br's outside of<br>
✔️ a code block
✔️ ```
✔️ This is OK\
✔️ inside a\
✔️ markdown code\
✔️ block
✔️ ```
✔️ `This is OK\`
````
Doing it this way will prevent `pandoc(1)` from generating deprecated
HTML elements such as `<br />`.
<div id="img"></div>
## Do not use markdown style images
Instead of using this markdown element to embed an image:
``` <!---markdown-->
![alt text](filename.png "Title") <=== no thank you
```<!---markdown-->
❌ ![alt text](filename.png "Title")
✔️ <img src="filename.png"
alt="describe the filename.png image for someone who cannot view it"
width=PIXEL_WIDTH height=PIXEL_HEIGHT>
Which allows for HTML <img> attributes `alt=`, `width=` and `length=`.
```
Use an `<img>` HTML element with `alt=`, `width=` and `length=`
attributes:
The problem is that `pandoc(1)` generates problematic HTML from the
markdown image construct, the resulting HTML does **NOT** have `width`
and `height` information so browsers have to slow down on rendering
text around the image until it can internally determine the image size.
``` <!---html-->
<img src="filename.png"
alt="describe the filename.png image for someone who cannot view it"
width=PIXEL_WIDTH height=PIXEL_HEIGHT>
```
<div id="hr"></div>
For example, instead of:
``` <!---markdown-->
![1984-anonymous-tattoo.jpg](1984-anonymous-tattoo.jpg) <=== no thank you
```
use this HTML:
``` <!---html-->
<img src="1984-anonymous-tattoo.jpg"
alt="image of a tattoo of the 1984 anonymous C code"
width=600 height=401>
```
The problem goes beyond the fact that `pandoc(1)` generates problematic
HTML from the markdown image construct, the resulting HTML does **NOT**
have `width` and `height` information so browsers have to slow down
on rendering text around the image until it can internally determine
the image size.
Jump to: [top](#)
<div id="hr">
<div id="horizontal">
<div id="lines">
## Please do NOT use markdown style horizontal lines
</div>
</div>
</div>
## Replace markdown style horizontal lines.
Please do **NOT** use `---` style lines in markdown to create horizontal
lines or to separate sections.
Unless something is inside a markdown **code block**, do **NOT** start a
line with 3 or more dashes ("`-`"s).
Unless something is inside a markdown **code block**, do **NOT** start
a line with 3 or more dashes ("`-`"s).
Such markdown causes `pandoc(1)` to generate `<hr />`. The `<hr />` has no
effect in standard HTML 5 and interacts badly with unquoted attribute values.
Such markdown causes `pandoc(1)` to generate `<hr />`. The `<hr />` has
no effect in standard HTML 5 and interacts badly with unquoted attribute
values.
If a horizontal line is really needed, use:
``` <!---html-->
<hr>
```<!---html-->
❌ * * *
❌ - - -
❌ ---
Regular divider...
✔️ <hr>
A short divider...
✔️ <hr style="width:10%;text-align:left;margin-left:0">
```
If a short line is needed, use:
<div id="parentheses"></div>
``` <!---html-->
<hr style="width:10%;text-align:left;margin-left:0">
## Do not put a literal "`(`" or "`)`" in markdown link titles
```<!---markdown-->
❌ [some (text)](https://example.com/curds)
✔️ [some &lpar;text&rpar;](https://example.com/curds)
❌ [ls(1)](https://example.com/ls-man-page.1)
✔️ [ls&lpar;1&rpar;](https://example.com/ls-man-page.1)
```
<div id="closing-parentheses"></div>
Jump to: [top](#)
<div id="parentheses">
## Please do NOT put a LITERAL "`(`" or "`)`" in markdown link titles
</div>
Please do **NOT** use literal parentheses inside markdown link titles.
Instead of:
``` <!---markdown-->
[some (text)](https://example.com/curds) <=== no thank you
```
use:
``` <!---markdown-->
[some &lpar;text&rpar;](https://example.com/curds)
```
Instead of:
``` <!---markdown-->
[ls(1)](https://example.com/ls-man-page.1) <=== no thank you
```
use:
``` <!---markdown-->
[ls&lpar;1&rpar;](https://example.com/ls-man-page.1)
```
Jump to: [top](#)
<div id="closing-parentheses">
## Please do NOT end markdown links with "`))`"
</div>
Please do **NOT** end a markdown link with a double closed parenthesis "`))`".
## Do not end markdown links with "`))`"
Markdown links that end in "`))`" complicate parsing and sometimes lead
to incorrect URLs or file paths.
Instead of:
```<!---markdown-->
❌ [some text](https://example.com/foo_(bar))
✔️ [some text](https://example.com/foo_&lpar;bar&rpar;)
``` <!---markdown-->
[some text](https://example.com/foo_(bar)) <=== no thank you
❌ This thing, ([some text](some/path)), is NOT ideal.
✔️ This thing, [some text](some/path), is better.
```
use:
<div id="italics"></div>
``` <!---markdown-->
[some text](https://example.com/foo_&lpar;bar&rpar;)
## Use `_` for italics in markdown
Please do **NOT** use `*` (single asterisk) for italics in markdown.
Instead use an underscore (`_`). Using an asterisk can complicate
parsing and sometimes lead to incorrect results. This can especially
go for when it is **_bold and italic_**.
```<!---markdown-->
❌ *this text is italic*
✔️ _this text is italic_
❌ ***this text is bold italic***
✔️ **_this text is bold italic_**
✔️ _**this text is bold italic**_
```
As another example, instead of:
``` <!---markdown-->
This thing, ([some text](some/path)), is NOT ideal. <=== no thank you
```
use:
``` <!---markdown-->
This thing, [some text](some/path), is better.
```
Jump to: [top](#)
<div id="code-text">
<div id="code-and-text">
<div id="text">
## Please do NOT place text on the IMMEDIATE (very next) line after a markdown code block
</div>
</div>
</div>
Please do **NOT** place text on the next line after a markdown code block.
Instead, place a blank line after the end of a markdown code block
as this makes it easier to detect when markdown code blocks are
**NOT** properly indented.
Instead of:
```` <!---markdown-->
```
int
main(int foo)
{
return foo;
}
```
C compilers cannot be given a -Wno-main-arg-errors flag. <=== no thank you
````
use:
```` <!---markdown-->
```
int
main(int foo)
{
return foo;
}
```
C compilers cannot be given a -Wno-main-arg-errors flag.
````
**BTW**: Please note the blank line after the code block.
Jump to: [top](#)
<div id="italics">
## Please use `_` for italics in markdown
</div>
Please do **NOT** use `*` (single asterisk) for italics in markdown. Instead use
an underscore (`_`). Using an asterisk can complicate parsing and sometimes lead
to incorrect results. This can especially go for when it is **_bold and
italic_**.
For example, instead of:
``` <!---markdown-->
*this text is italic* <=== no thank you
```
use:
``` <!---markdown-->
_this text is italic_
```
Another example, for **_bold italic_**:
Do **NOT** use:
``` <!---markdown-->
***this text is bold italic*** <=== no thank you
```
Instead use:
``` <!---markdown-->
**_this text is bold italic_**
```
or:
``` <!---markdown-->
_**this text is bold italic**_
```
**BTW**: the astute reader might notice that some cases of `*` for italic might
have slipped through. We do ask you, however, to please observe this rule, as
best you can.
**BTW**: the astute reader might notice that some cases of `*` for
italic might have slipped through. We do ask you, however, to please
observe this rule, as best you can.