Wow, and I was going to finish my python port this morning. Thanks for saving me the time.
Cool, I was just working on this a couple of days ago but I’m glad someone else beat me to it. :) I think I’ll offer it as a text formatting option in the next release of my multiprotocol Python XML-RPC blog client/library…
Excellent!
Now I guess it’s just a matter of Gruber making an Applescript version and the Narrator doing it up in Java und Textile vill rule ze vorlt.
Haha!
(Whoops, war on)
Anyone know whether someone has written a side-by-side or feature-by-feature comparison of Textile with reStructured Text? And/or with the “HTML-Lite” that Radio UserLand implements?
reStructured Text is described in these two spots:
http://docutils.sourceforge.net/spec/rst/introduction.html
http://docutils.sourceforge.net/docs/rst/quickref.html
— Doug L. 
Quickly:
reStructuredText is designed to be output-format-agnostic; Textile is specifically designed for web display.
reStructuredText has many options that Textile lacks, such as field lists, definition lists, nested blockquotes, simple tables, and grid tables.
Textile has several options that reStructuredText lacks, such as inline shorthand for CODE, CITE, and SUP. Also, its ability to auto-map high-bit ASCII characters to HTML numeric entities (making it possible to cut-and-paste smart quotes without breaking validation).
reStructuredText is designed for a particular style of formal document: it has a system of cataloging footnotes, bibliographic citations, and hyperlinks. Textile is designed for shorter, more informal documents like blog posts and comments.
— Mark
Is this doable in JavaScript? I’d like to see this without a screen refresh.
— pb
I don’t see why not. Javascript has regular expressions.
— Mark
So Dean … are you giving me your permission?
Permission granted, Dan. Encouraged, cheered on, supported, cheer-led, voted for, acknowledged, endorsed, sanctioned (whups), say hey, and fire up a Cohiba.
So it’s time, then, for textile() to have a plain-language spec: doctypes for which it intends validity and what it generally aims to be. Again (I keep saying this), a description of what it isn’t probably does the job best.
What is it? Structurally-aware UBBcode with an addiction to Proust, validating to XHTML 1.1 Trans. The earth will shift under its feet, though. To live un-Balkanized, it must define itself.
Nobody commenting where would implement a half-Textile, a sorta-Textile — but others might. While it sill can, I think Textile should announce its goals, benefits and understood limitations [somewhat] more formally.
My Mom will like Txl because it’s intuitive, but in two years when she wants a PDA version of her etchings she’ll thank me because it delivered valid, transformable structure without her knowing it. These are its twin gifts.
Mozilla and IE now feature API’s for rich text editing. It would be pretty fly if you integrated PyTextile with that.
-Ken
Ken: Do you know any more info on that? It sounds like a fun project to work on.
pb: If you’re going to work on that, please let me know; I’d love to participate.
Mark: see pb’s comment — you should probably have code to convert an (obvious) email address into the mailto: link instead of a broken http:// one. I can help with this if you don’t have time.
Um, the box is clearly labeled “Home page”. I don’t want to be responsible for anyone’s email address.
— Mark
Mark: OK, fair enough. But, as you know, people don’t always read directions, much less follow them. :)
Hmm, something goes wrong with consecutive lists. For example, try:
* one
* two
* three
* four
* five
This should yield two separate lists, but instead PyTextile creates one list. Using a *-list directly after a #-list doesn’t work well either.
— Hans
1.04 is up and should fix the problem with consecutive lists.
— Mark
Nice work! One more contraction for you, that I’d added in my local copy (in addition to ‘ve): ‘re (as in “they’re”; that may be the only instance in English).
It doesn’t seem to have appeared so far, so here’s a Java version (with free bugs!)