Markdown development

[Aristotle Pagaltzis]

* David E. Wheeler <david at kineticode.com> [2010-03-20 23:05]:

> I'm pretty happy with MMD definition lists (with s/:/~/g)


I tried to like them.


> and with psql-style tables (mostly implemented in

> Markdent::Dialect::Theory. I find them both esthetically

> pleasing.


They look OK but are a pain to edit. (I haven’t seen a table
syntax that I find easier to edit than raw HTML. Well, the fact
that Markdown won’t touch the contents makes them still a pain.)

I think this editing|reading tension is fundamental, therefor not
resolvable.


* Sherwood Botsford <sgbotsford at gmail.com> [2010-03-20 23:35]:

> G. K. Chesterton commented, "If something is worth doing, it's

> worth doing badly"


Only within limits.


> I like MMD's table syntax. Not perfect. Still a pain to

> construct, especially if you want to keep the notion of having

> to look reasonable as plain text.  But it *really* beats

> <table><th><td>...</th><tr><td></td>...</tr>...</table>


To look at? Beats it. To write? Not hardly.

Esp. if you leave off the closing tags, as I do these days when
I write tables in Markdown documents.


> At this point I can do it with template toolkit and include

> files, but it's more than a bit rube-goldbergish. In the league

> with programmable candle powered hydralic peanut butter

> spreaders.


Actually include files are how I would suggest you do that.
Better yet use a Markdown implementation where you can pass
a prepopulated table of link references (so that the Markdown
formatting won’t have to parse the same 1,400 link references
over and over).


> Easy way to modify the behaviour with certain tags.


One of the things I want is some way to make it easier to tell
Markdown to re-engage inside block tags, in a less klunky fashion
than Markdown Extra’s (I think?) markdown=1 pseudo-attribute.
I don’t have a good idea for this, though.

I wouldn’t want any configurable behaviours, though. To me it is
a very important point that you can take a Markdown document from
one environment to another without breakage.


* Seumas Mac Uilleachan <seumas at idirect.ca> [2010-03-21 00:45]:

> I hear constantly about needing "Gruber's blessing" for any

> overhaul or changes to Markdown. Why?


In case you are referring to my own recent mail, I never said it
needs *Gruber*’s blessing specifically. What it does need is one
central voice. If you think it doesn’t, then ask yourself why all
of the reimplementations have added their own features, yet none
of them have copied each other.

Except that several of them have copied Markdown Extra. As it
happens, Michel Fortin had some blessing from Gruber on several
of his efforts. Coincidence?

No one needs permission or blessing to fork Markdown. Many people
have done it.

But that’s the point. To get more than Just Another Fork, it neds
to have weight of voice to bring all the other forks together
behind it.


> The goal of markdown is readability. There is no such thing as

> a readable html table. I would argue that tables are a useful

> enough feature to include. Whether it is done badly or well is

> often subjective. At the minimum a simple table format would be

> important to me (not requiring spanning cells or complex table

> layouts). Tables are the easiest way to list corresponding

> values or data that they really should be somehow included.


The problem with tables as I see it is as above: I think that
tables fundamentally cannot be both easy to edit and easy to read
within the constraints of plaintext.

Regards,
-- 
Aristotle Pagaltzis // <http://plasmasturm.org/>