New link syntax

[Lou Quillio]

On Mar 31, 2004, at 2:42 PM, european bob wrote:

> On Wed, 2004-03-31 at 20:16, Lou Quillio wrote:
>> Understand.  What if a user *intends* for the duplicate link text to
>> point to the same link target?
>
> If they intended that, they wouldn't redefine the link target.

Right.  Got ahead of myself again.  I'm changing my position, and 
oppose implicit links.  Here's why.

Two ways to build a link:  (1) explicit inline, and (2) explicit 
referential.

Explicit inline offers no real advantages over inline XHTML markup.  
Maybe it's a little more readable, but you lose height/width attributes 
-- and that inline URI causes the same textarea horizontal-scroll 
problem as inline XHTML (this is true for links and images).  The other 
advantage of inline MD link syntax is that you stay in the MD paradigm 
of square brackets.  The negatives remain.  The upshot is that MD has 
to offer inline, but it's no great shakes.

Explicit referential links smoke.  They get those (generally) 
unreadable URIs out of the body, defeat horizontal scroll in textareas, 
and are reusable.  They make clear that there's a linked resource but 
treat it like a footnote you can consult separately, elsewhere in the 
text blob, without breaking thought.  Explicit referential links are 
some serious shit imo, and referential image tags only make things 
better.  Let's not forget that these are all in-house considerations:  
the output source is inline XHTML, and that's all a consumer sees.  
Internal workflow is the issue.

As Euro-Bob rightly says, implicit links can undercut the combination 
of explicit referential *and* explicit inline links.  I'd think most 
users would choose between inline and referential and stick to one or 
the other, but they might not, confusing link-target/link-anchor 
relationships.  Either these three can't coexist, or implicit links 
have to be an adjunct, subject to rules.

To my mind, if implicit links are done they should be adjuncts to 
explicit referential links elsewhere declared, not to explicit inline 
links elsewhere declared.

They can't be both.  That's the problem.

Considering that referential links are easy -- almost natural -- to 
repeat if needed, the advantage of implicit links is all but vitiated 
in referential-link authoring practice.  Only explicit-inline practice 
has any use for them.

So implicit links should attach to explicit inline links elsewhere 
declared, if implicit links are allowed at all.  I'm thinking they 
shouldn't be, because they raise conflict resolution issues not 
*essential* to advancing the project.

Why do users of explicit MD inline links -- who might as well be 
writing inline XHTML, anyhow -- need their own shorthand that jacks 
another vector into MD?  The shorthand already exists:  referential 
linking.  Use it.  Hi.

LQ