Markdown Discuss - Apr 2011 - ol start with a specific number?

We recently received a request and working patch for Python-Markdown
which adds support for starting an ordered list with the number given
on the first line. For example:

    3. Foo
    4. Bar

would result in

    <ol>
    <li start="3">Foo</li>
    <li>Bar</li>
    </ol>

I'm not opposed to adding this, but I noticed that no other
implementation (of those on Babelmark) implements this by default (not
counting Pandoc's extended mode). I haven't checked if other
implementations offer this as an option.

My question is: should this be an option to turn on and off, and if
so, should it be on or off by default?

JG states in the docs:
> If you do use lazy list numbering, however, you should still start
> the list with the number 1. At some point in the future, Markdown
> may support starting ordered lists at an arbitrary number.
Given that statement, it would seem that on by default and without an
option to turn if off would be fine. But what is the reality in the
real world? If I did that, would a bunch of documents suddenly start
rendering incorrectly - or at least different that expected? I guess
the real question is: has everyone been ignoring that piece of advice
in the docs and if so, is this something we should care about as
implementors?

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

* Waylan Limberg <waylan at gmail.com> [2011-04-07
04:15]:
> I'm not opposed to adding this, but I noticed that no other
> implementation (of those on Babelmark) implements this by
> default (not counting Pandoc's extended mode). I haven't
> checked if other implementations offer this as an option.
John?s reason was that the `start` attribute was deprecated
in HTML?4 Strict. He has since said at least once that this
reasoning was flawed and the decision a mistake.
> My question is: should this be an option to turn on and off,
> and if so, should it be on or off by default?
Bitter as it may be I think you would need to leave it off by
default. Interoperability is important.
> Given that statement, it would seem that on by default and
> without an option to turn if off would be fine.
I wish John would make another release to straighten out these
handful of known tiny niggles? since his is the implementation
that everyone else?s will follow.
> But what is the reality in the real world? If I did that, would
> a bunch of documents suddenly start rendering incorrectly - or
> at least different that expected? I guess the real question is:
> has everyone been ignoring that piece of advice in the docs and
> if so, is this something we should care about as implementors?
It?s not that simple. Consider what happens if some user writes
a document while previewing it with your implementation, which
advertises itself as Markdown, then pastes it into the textarea
in some web app, which also advertises Markdown support.

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