list corner case

[Waylan Limberg]

On Tue, Sep 9, 2008 at 5:58 PM, Tom Humiston <tom at jumpingrock.net> wrote:

> Shorter version of my previous post:

>

> Gruber gives this example...

>

> 3. Bird

> 1. McHale

> 8. Parish

>

> ...and states that it will be numbered 1-2-3. Which is enough to make clear

> that in Markdown's design, the kind of goofy content in list-item markers

> that we're discussing is ignored. Simple.

>


In other words, you're arguing for option A.

However, while, the above example indicates what the expected behavior
is when the list indicators are of the same type, it provides no hints
as to what should be expected when the list indicators are of
different types. You seem to be suggesting that we shouldn't even look
at that, but we have to to identify that line as either (1) another
list item of the same level, or (2) an additional line of the previous
list item.

That second possibility is what you left out of your long post, and
what forces every implementation to specifically look for the list
indicator on every item - not just the first. In my observation, most
implementations use different code (usually different regex) to check
for ol indicators than for ul indicators. So (nearly) every
implementation is going to notice the difference in indicator style,
while they may not notice the inconsistency in numbering of a numbered
list (that's the nature of regex).

Those that follow option A have consciously chose to ignore that
difference in type. The thing is, humans do not ignore that
difference. I realize most humans do not ignore the non-sequential
numbering either and that's why I'm willing to settle for option A as
a close second to option C - the way my brain sees it.


-- 
----
Waylan Limberg
waylan at gmail.com