Ordered list syntax.

[Lou Quillio]

On Mar 28, 2004, at 12:05 AM, Jelks Cabaniss wrote:

> The W3C Validator can't use the catalog mechanism when it doesn't 
> recognize
> the "Formal Public Identifier".  And that's precisely what happened 
> when it
> saw "1.1 Transitional", because no such beast exists.

> Try it!  :)

I just tried it.  Shit.

Whoa.  How long has it been since I followed that W3C validator link 
that says "The Validator XML support has some limitations"?  Too long, 
I guess.  Your (Jelks') informative examples certainly point to the 
same DTD.  Why does the validator report success/failure in terms of my 
named doctype rather than in terms specific to the DTD I cited.  The 
validator output shouldn't compound the fact that I'm a dope.

> Anyway, DTDs are for validation purposes only. [et al.]

Right, but user-agent rendering increasingly keys on the declared 
doctype, no?  The idea is to (1) compose in MD and (2) feel sure that 
MD's XHTML transformation won't freak user-agents.  Obviously 
user-agents are uncontrollable third actors.  My fear was that common 
user-agents will soon begin to mis-render lists with the "deprecated" 
attributes (perhaps because of declared doctype), be they Web browsers 
or WAP phones or whatever.  We want our source documents to have long 
and practical lives.  If the list-control wind will soon blow from a 
different direction -- in terms of how user-agents will implement the 
standards -- where's the safest place to be going forward?

Split-ordered lists are an odd case, to be sure.  Almost nobody needs 
them, including most of the somebodies who think they do:  probably 
there's a better way to organize the information.  J.G. said he wanted 
to accommodate them, and it seemed important to me that the XHTML 
output be as durable as possible.

So that's the question.  It's a judgment, a forecast, a guess.  What's 
the most durable custom-list markup, and what do authors need to know 
about the transformative assumptions that MD makes?  UA rendering 
matters, and UAs make rendering judgments on document conditions 
external to the (transformed) MD text-blob.

> Or use your *own* DTD!

Umm, dude?  Dude?  Who's gonna do that?  Doesn't that trail lead to 
everybody having their own pseudomarkup, their own parsers, and their 
own DTDs to make sense of them?  Not a bad place at all, but I think 
MD's goals are more modest.

Unless ... unless ... unless maybe that's what's short-circuited these 
things heretofore, the missing piece.  A public Markdown DTD or some 
such, to finish things off.  Assuming user-agents eventually take 
rendering cues from DTDs or CMSes challenge content against them.  
Okay, I have to stop thinking now.

> 2.  Allow for MD equivalents of 'start' and 'type', and let the user be
> responsible for his or her own DOCTYPE declaration[s].

But there it is again, that big-think roll-and-own-your-own XSLT thing. 
  Oy.  I need some Excedrine RDF.  Maybe watching television will help.

LQ


++++++++++++++++++++++++++++++++++++
Lou Quillio
P.O. Box 24
Saratoga Springs, NY, USA 12866
518.796.0256 (cell)
http://quillio.com/