multiline header

[Waylan Limberg]

On Tue, Mar 2, 2010 at 11:55 AM, david parsons <orc at pell.chi.il.us> wrote:

> In article <e5d795bb1003020800x4b755d7fu5af6be6b4ea685d2 at mail.gmail.com>,

> Waylan Limberg  <markdown-discuss at six.pairlist.net> wrote:

>

>>The original design goals also specifically stated that anything more

>>complex that the syntax already supported should be handled by raw

>>html. And, well, if you look at the source of his pages (add ".txt" to

>>the urls), J.G. almost exclusively uses raw html for his headers.

>

>    Hmm?   From a casual look at Daring Fireball this morning, about

>    the only raw html headers I saw were ones where he was id'ing

>    them (since, alas, the standard doesn't have the [foo](id:bar)

>    pseudo-protocol.)


Figures. I almost went and double checked before writing that, but
went off of memory alone. Memory tends to exaggerate things like that.
That's what I get.


>

>>But

>>that's not what I would call designed for lazy users.

>

>    It might be just me, but the surprise factor of having a header

>    reach back and grab an entire paragraphs might be less attractive

>    than it would seem, even to the body of users who write long headers

>    with a text editor that forces line-wrap at 80 characters.

>


You have a point. However, I tend to almost exclusively use hash
headers (less typing) and as the first line of the header is always
defined, it wouldn't be a problem. The problem is that the paragraph
after could get consumed if the author failed to include a blank line.

To be honest, I can't think of any elegant way to do multi-line setext
headers, but multi-line hash headers should be easy.


-- 
----
\X/ /-\ `/ |_ /-\ |\|
Waylan Limberg