Adding additional output formats for Markdown?

[Fletcher T. Penney]

If this response is a bit long, it's because I actually have a day  
off tomorrow, and had time to think about this tonight.  It is not  
intended as a rant.  ;)


On Jul 27, 2005, at 9:55 AM, John Gruber wrote:


>

> I think you're mixing apples and oranges here. You're building both

> an extension to the Markdown syntax *and* you're building a

> publishing tool.

>

> It's a fine idea to build a publishing tool based on plain text

> files, where each file has a defined structure of one or more

> metadata headers at the beginning, followed by a blank line,

> followed by Markdown-formatted text (or, in your case,

> Markdown-plus-extra-stuff-formatted text).

>

> But I think it's a mistake, or at least an unnecessary conflation,

> to include the headers as part of your syntax itself. The

> Markdown-ish syntax should only apply to the prose itself.

>

> Most of my use of Markdown is in Movable Type, but there's a reason

> why there's nothing MT-specific in the Markdown syntax itself.

> Whereas in your project, the syntax is implicitly tied to your

> specific publishing model.


I disagree (or perhaps we are looking at it differently).  I am  
trying to do the opposite.  Every (almost) use of Markdown involves  
some form of explicit or implicit metadata - Movable Type documents  
have a title (url), they have an author, they have a date of  
publication, modification, etc.  Some of this data exists, but cannot  
be accessed, and some of it is understood, if not actually defined  
(you are understood to be the author of  Markdown formatted page on  
your Movable Type website, even if it is not shown on the page  
somewhere).

Similarly, if you reformatted a Markdown text file as a pdf, the  
author would be the same, but perhaps the title you would want might  
be different.  Or the date of publication.  Or the url.

Blosxom specifically has a meta plugin to handle this sort of  
information.  Movable type as the MT-Meta plugin.  They are not  
intrinsically compatible (I believe - I haven't used blosxom in a  
year or two, and don't use MT)

For me, my dream for Markdown would be to use it as the default  
syntax for most of my text-ish documents - weblogs, wiki pages,  
academic papers, readme's for my software, etc.  A single Markdown  
document would likely end up becoming a variety of different kinds of  
documents, each of which would have a different means of specifying  
the title, author, etc.

By defining a standard way of (optionally) including that information  
within the document (in much the same way that a printed email, or a  
piece of business correspondence includes the sender, recipient,  
date, and possibly a subject), you can the use that information in  
whatever way is appropriate for a given output format.

The example document on my website includes metadata for author,  
title, date, and keywords.  All of the data is used to form the  
metadata for the pdf.  The title is used to create the title in the  
XHTML header for the XHTML version.  The utility of this is self- 
evident.

By "stick[ing] with the Markdown philosophy", I mean that the reason  
I continue to choose Markdown over other syntaxes is that I like the  
way it looks.  It is easy to remember, and easy to interpret when you  
are reading "raw source."  I feel like the example of metadata I gave  
is also inherently readable.  Any individual implementation of a  
Markdown application is free to ignore it ( I understand that it is  
not, and may not ever be a part of the official Markdown syntax, and  
therefore would not exist in order to be parsed out. ;)  But for  
"Markdownish" programs that use it, many new opportunities are opened  
up.

The headers would not be interpreted by the Markdown Parser in any  
way, if that is what you meant by:


> The Markdown-ish syntax should only apply to the prose itself.


I do not envision lists, or code blocks, etc within the metadata.   
Straight text only.

As for:


> Whereas in your project, the syntax is implicitly tied to your

> specific publishing model.


That is exactly what I am trying to free Markdown from.  Currently,  
you cannot use a single Markdown text file to generate a LaTeX pdf  
with a title page _AND_ an XHTML web page with a title without human  
intervention.  With my program, you can do this, and much more.  The  
metadata, and the rest of the document, can be used for any type of  
document creation one would like.  And if (big if) the meta-data was  
part of the official syntax, then other tools would be free to ignore  
it.  In any case, no one is required to include metadata in the  
source document.

Markdown was originally designed (per your website) as a tool to  
assist in crafting valid (X)HTML.  I understand this.  But the syntax  
could be used for much more, and from searching the archives, I am  
not alone in this desire or in creating tools to try and make it  
reality.  I really like the pdf format, and would like to be able to  
type a file in "Markdown-ese", and have it automatically converted to  
a pdf using my favorite LaTeX style.  With MultiMarkdown (or whatever  
it should be called), this is a trivial problem.  A single shell  
script can convert a single Markdown text file to an unlimited  
variety of output formats without any human intervention.

A weblog entry could be downloaded as a pdf, or an ebook, or an RTF,  
or whatever.  "The possibilities are endless..."  ;)


>> Please let me know what you think.  (Also, John (and anybody else) if

>> I have incorrectly used any code, or stepped on toes - please let me

>> know!!!)

>>

>

> What you've done is perfectly acceptable given Markdown's licensing.

> I don't really think it's a great idea to fork the formatting syntax

> in the ways you're proposing, but I don't have a problem with it

> either.

>

> -J.G.


Like I imagine others (eg Michel Fortin) would like, I would love to  
have a single standard version of Markdown that incorporates these  
features, just as I would like to have standardized support for  
tables.  My goal is not to define "the" format for metadata.  Nor to  
insist that you have to change your vision of Markdown.

It is simply to point out that with one simple addition, the utility  
of Markdown is greatly enhanced.  And, yes, I am aware of the trap of  
incremental feature bloat and love Markdown's simplicity.  If no one  
else out there wants to use these features, then I can easily  
continue to hack my own copy of Markdown and use it as desired.  But  
I wanted to bring the example up for demonstration and discussion in  
case others were interested.

And as always, feedback is welcome!

Thanks to anyone who actually read this far,

Fletcher


-- 
Fletcher T. Penney
fletcher at alumni.duke.edu

An intelligent man is sometimes forced to be drunk to spend
time with his friends.
         - Ernest Hemingway


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 3949 bytes
Desc: not available
Url : http://six.pairlist.net/pipermail/markdown-discuss/attachments/20050727/603dd648/smime-0001.bin