On simplifying table syntax in any future markdown extension. (Use CSV)

Sean Leonard dev+ietf at seantek.com
Fri Sep 5 12:47:52 EDT 2014


On 9/5/2014 6:38 AM, Michel Fortin wrote:
> Le 5-sept.-2014 à 9:13, mofo syne <mofosyne at gmail.com> a écrit :
>
>> Btw csv is a pretty loose standard. So the above csv should work still
>> for many csv parsers.
> Which just add another problem. If people come to expect CSV to work for Markdown tables, they'll copy-paste their CSV documents generated by their spreadsheet application and expect it to work, which it won't because it just supports a small subset of CSV that isn't what most application generate. Then I'll get a deluge of bug reports about this or that getting mangled and all those bugs be unfixable because CSV is a too loose standard to parse it reliably. Better stay away from CSV.

I agree with this.

There is an informal standard: RFC 4180, "Common Format and MIME Type 
for Comma-Separated Values (CSV) Files" 
<http://tools.ietf.org/html/rfc4180>. It is informal because Section 2 
describes the way most people do it, rather than prescribing rigid 
adherence.

If you want to go CSV -> Markdown, your best bet is to use a CSV -> HTML 
processor, and then insert the HTML into the Markdown. Unambiguous. 
Problem solved.

Markdown tables should be as simple as possible...and should look like 
the desired HTML output, in plain text e-mail. For this purpose, the 
pipe format works all right.

-Sean



More information about the Markdown-Discuss mailing list